Drug-Stability - PDFCOFFEE.COM (2024)

This Page Intentionally Left Blank

James Swarbrick, AA], Inc. ~i~mington, North Carolina

Larry L. Augsburger University of Maryland Baltimore, Maryland Douwe D. Breimer Gorlaeus Laboratories Leiden, The Netherlands

David E. Nichols Purdue University West Lafayette, Indiana Stephen G. Schulman University of Florida Gainesville, Florida

Trevor M. Jones The Association of the British Pharmaceutical Industry London, United Kingdom

Jerome P. Skelly Alexandria, Virginia

Hans E. Junginger Leiden/Amsterdam Center for Drug Research Leiden, The Netherlands

Felix Theeuwes Alza Corporation Palo Alto, California

Vincent H. L. Lee University of Southern California Los Angeles, California

Geoffrey T. Tucker University of Sheffield Royal Halla~shireHospital Sheffield, United Kingdom

Peter G. Welljng lnstitut de Recherche Jouveinal Fresnes, France

A Series of Textbooks and Monographs

edited by

AAI, Inc. ~ i l ~ i n g t a~a~~ n, Carolina

1 . Pharmacokinetics, Milo Gibaldi and Donald Perrier 2. Good Manufacturing Practices for Pharmaceuticals: A Plan for Total Quality Control, Sidney S. Willig, Murray M. Tuckerman, and ~ i l l i a mS. Sitchings IV

3. ~icroencapsuiation, edited by J. R. Nixon 4. DrugMetabolism:ChemicalandBiochemicalAspects, BernardTestaand Peter Jenner 5. New Drugs: Discovery and Development, edited by Alan A. Rubin 6. Sustained and Controlled Release Drug Delivery Systems, edited by Joseph R. Robinson odern Pharmaceuti~s, edited by Gilbert S. Banker and C~ristopher T. Rhodes 8. Prescription Drugs in Short Supply: Case Histories, Michael A. Schwartz 9. Activated Charcoal: Antidotal and Other Medical Uses, David 0, Cooney 10. Concepts in Drug ~ e t a ~ o l i s(in m t w o parts), edited by PeterJennerand Be~nardTesta 1 1, Ph~rmaceutical Analysis: Modern Methods (in t w o parts), edited by James W. Munson 12. Techniques of Solubilization of Drugs, edited by Samuel S. Y a l ~ o ~ s k y 13. Orphan Drugs, edited by Fred €. Karch 14. Novel Drug Delivery Systems: Fundamentals, Developmental Concepts, Biomedical Assessments, We W' Chien 15. ~harmacokinetics: SecondEdition,Revisedand Expanded, Milo Gibaldiand Don 16. Goo turing Practices for Pharmaceuticals: A Plan for Total Quality Control, Second Edition, Revised and Expanded, Sidney S,Willig, Murray M, Tuckerman, and William S. ~itchings /V 17. Formulation of Veterinary Dosage Forms, ~ d i t e dby Jack Blodinger 18. Dermatological Formu~ations:Percutaneous Absorption, Brian W. Barry 19. The Clinical Research Process in the Pharmaceutical Industry, edited by Gary M. Matoren 20. Microencapsulation and Related Drug Processes, Patrick B. Deasy 21, Drugs and Nutrients: The Interactive Effects, edited by D a p ~ n eA. Roe and T. Colin Cam~bell 22. Biotechnology of Industrial Antibiotics, €rick J. Vandamme 23. Pharmaceutical Process Validation, edited by Bernard T. Loftus and Robert A, Nash

24. AnticancerandInterferonAgents:Synthesisand

Properties, editedby

Raphael M. Ottenbrite and George B. Butler

25. Pharmaceutical Statistics: Practical and Clinical Applications, Sanford Bolton 26. Drug Dynamics for Analytical, Clinical, and Biological Chemists, enj jam in J. Gudzinowicz, Burrows T. Younkin, Jr., and Michael J. Gudzinowicz 27. Modern Analysis of Antibiotics, edited by AdjoranAszalos 28. Solubility and Related Properties, Kenneth C. James 29. ControlledDrugDelivery:FundamentalsandApplications,Second Edition, Revised and Expanded, edited byJoseph R. Robinson and Vincent H. Lee 30. New Drug Approval Process: Clinical and Regulatory Management, edited by Richard A. Guarino

31 . Transdermal Controlled Systemic Medications, edited by Yie W. Chien 32. DrugDeliveryDevices:FundamentalsandApplications, editedbyPraveen Tyle

33. Pharmacokinetics: Regulatory * Industrial * Academic Perspectives, edited by Peter G. Welling and Francis L. S. Tse 34. Clinical Drug Trials and Tribulations, edited by Allen E: Cat0 35. TransdermalDrugDelivery:Developmental Issues and Research Initiatives, H, Guy edited by Jonathan Hadgraft and Richard

36. AqueousPolymericCoatings forpharmaceutical DosageForms, editedby James W. McGinity 37. Pharmaceutical Pelletization Technology, edited by lsaac Ghebre-Sellassie 38. Good Laboratory Practice Regulations, edited by Allen F. Hirsch 39. Nasal Systemic Drug Delivery, Yie W Chien, Kenneth S. E. Su, and Shyi-Feu Chang 40. Modern Pharmaceutics:Second Edition,Revised and Expanded, editedby Gilbert S. Banker and ~hristopherT. Rhodes 41. Specialized Drug Delivery Systems: Manufacturing and Production Technology, edited by Praveen Tyle 42. Topical Drug Delivery Formulations, edited by David W. Osborne and Anton H. Amann 43. Drug Stability: Principles and Practices, Jens 7. Carstensen 44. Pharmaceutical Statistics: Practical and Clinical Applications, Second Edition, Revised and Expanded, Sanford Bolton edited by Mark Chasin 45. Biodegradable Polymers as DrugDeliverySystems, and Robert Langer 46. PreclinicalDrugDisposition: A LaboratoryHandbook, Francis L. S. Tse and James J, Jaffe 47. HPLC in the Pharmaceutical Industry, edited by Godwin W, Fong and Stanley K. Lam 48. Pharmaceutical Bioequivalence, edited by Peter G. Welling, Francis L.S. Tse, and Shrikant V. Dinghe 49. Pharmaceutical Dissolution Testing, Urnesh V. Banakar 50. NovelDrugDeliverySystems:Second Edition,Revised and Expanded, Yie W. Chien 5 1 . Managing the Clinical Drug Development Process, David M. Cocchetto and Ronald V. Nardi 52. Good Manufacturing Practices for Pharmaceuticals: A Plan for Total Quality Control, Third Edition, edited by SidneyH. Wiilig and JamesR. Stoker 53. Prodrugs: Topical and Ocular Drug Delivery, edited by Kenneth B. Sloan 54. Pharmaceutical Inhalation Aerosol Technology, edited by Anthony J. Hickey

editedbyAdrianD.

55. Radiopharmaceuticals:ChemistryandPharmacology, Nunn

56. New Drug Approval

Process: Second Edition,Revised and Expanded, edited

by RichardA. Guarino

57. Pharmaceutical Process Validation:Second

Edition,Revised

and Expanded,

edited byIra R. Berry and Robert A. Nash 58. Ophthalmic Drug Delivery Systems, edited by Ashim K. Mitra 59. Pharmaceutical Skin Penetration Enhancement, edited by Kenneth A. Walters and Jonathan Hadgraft 60. Colonic Drug Absorption and Metabolism, edited by Peter R. Bieck 61 Pharmaceutical Particulate Carriers: Therapeutic Applications, edited by Alain Rolland 62. Drug Permeation Enhancement: Theory and Applications, edited by Dean S. Hsieh 63. Glycopeptide Antibiotics, edited by Ramakrishnan Nagarajan 64. Achieving Sterility in Medical and Pharmaceutical Products, Nigel A . Halls 65. Multiparticulate Oral Drug Delivery, edited by Isaac Ghebre-Sellassie 66. Colloidal Drug Delivery Systems, edited by Jorg Kreuter 67. Pharmacokinetics:Regulatory * Industrial 0 Academic Perspectives, Second L. Tse Edition, edited by Peter G. Welling and Francis S. 68. Drug Stability: Principles and Practices, Second Edition, Revised and Expanded, Jens T. Carstensen

69. Good Laboratory Practice Regulations: Second Edition, Revised Expanded, edited by Sandy Weinberg 70. PhysicalCharacterization of Pharmaceutical Solids, editedbyHarry

and

Brittain

71. Pharmaceutical Powder Compaction Technology,

edited by Goran Alderborn

and Christer Nystrom

72.Modern

Pharmaceutics:Third

Edition,Revised

and Expanded, editedby

Gilbert S. Banker and ChristopherT. Rhodes

73. Microencapsulation: Methods andIndustrialApplications,

edited by Simon

Benita

74. Oral Mucosal Drug Delivery, edited by Michael J .Rathbone 75. ClinicalResearch in PharmaceuticalDevelopment, edited by Barry Bleidt and Michael Montagne

76. The Drug Development Process: Increasing Efficiency and CostEffectiveness, editedbyPeterG.Welling,LouisLasagna,andUmesh

Banakar

77. MicroparticulateSystemsfor

the Delivery of Proteinsand Vaccines, edited

by Smadar Cohen andHoward Bernstein

78. Good Manufacturing Practices for Pharmaceuticals: A Plan for Total Quality Control, Fourth Edition,Revised and Expanded, Sidney H, Willig and James R. Stoker

79. Aqueous Polymeric Coatings for Pharmaceutical Dosage Forms: Second Edition, Revised and Expanded, edited by JamesW. McGinity 80. PharmaceuticalStatistics:PracticalandClinicalApplications,Third Edition, Sanford Bolton

81 . HandbookofPharmaceuticalGranulationTechnology,

editedbyDilip

Parikh

82. Biotechnology of Antibiotics: by William R. Strohl

Second Edition,Revised

and Expanded, edited

83. Mechanisms of TransdermalDrugDelivery, edited by Russell 0. Potts and Richard H- Guy 84. Pharmaceutical Enzymes, edited by Albert lauwers and Simon Scharpe 85. Development of Biopharmaceutical Parenteral Dosage Forms, edited by John A. ont tempo 86. Pharmaceutical Project Management, edited by Tony ~ennedy 87. DrugProducts for ClinicalTrials: An InternationalGuide t o Formulation * Production 0 Quality Control, edited by Donald C, Monkhouse and Christopher T. Rhodes 88. DevelopmentandFormulation of VeterinaryDosageForms:SecondEdition, J. Desmond 5aggot Revised and Expanded, edited by Gregory E. Hardee and 89. Receptor-Based Drug Design, edited by PaulLeff 90.Automation andValidation of Informationin PharmaceuticalProcessing, edited byJoseph F. deSpautz edited by Michael S. ~ o b e r t s 91. Dermal Absorption and Toxicity Assessment, and KennethA. Waiters 92. ~harmaceutical Experimental Design, Gareth A. Lewis, Didier ~ath;eu,and Roger Phan-Tan-luu

93. Preparing for FDA Pre-Approval Inspections, edited by Martin D. Hynes lllv 94. Pharmaceutical Excipients: Characterization by IR, Raman, and NMR Spectroscopy, David E. Bugay andW. Paul F;ndlay 95. Polymorphism in Pharma~euticalSolids, edited by Harry G.~rittain 96. Freeze-~~in~lLyop~il~zation of ~harmaceuticaland B i o l o ~ i c a l ~ r o d u cet d ~ ,i ~ e ~ by Louis Rey and JoanC. ay 97. er cutaneous Absorption: ~ r u g s - ~ o s m e t i c ~ e c h a n i s m s ~ ~ t h o ~ o l o g y , ThirdEdition,Revisedand Expanded, edited by Robert L. 5ronaug~and Howard 1. Ma;bach 98. Bioadhesive Drug Delivery Systems: ~undamentals,Novel Approaches, and Development, editedbyEdith Mathiow;tz, ~ o n a l dE. chick er in^ Ill, and Claus-Michael L ehr 99. Protein Formulation and Delivery, edited by Eugene J . ~ c ~ a l l y 100. New Drug Approval Process: Third Edition: The Global Challenge, e ~ i t e by d Richard A , Guarino 101. Peptide and Protein Drug Analysis, edited by Ronald E. Reid 102. TransportProcesses in PharmaceuticalSystems, ed;tedby Gordon Amidon, Ping 1. lee, and ElizabethM. Topp 103. Excipient Toxicity and Safety, edited byMyra L. Weiner andlois A. Ko~kosk;e 104. TheClinicalAudit in pharmaceuticalDevelopment, ed;tedby M i c ~ a eR. l Hamrell

105. PharmaceuticalEmulsionsandSuspensions,

edited by Francoise ~ i e / / o ~ d

and ~ilberte Marti-Mestres

106. OralDrugAbsorption:PredictionandAssessment,

editedby

Dressman and Hans Lennernas

107. Drug Stability: Principles and Practices, Third Edition, Expanded, ed;ted by Jens T. Carstensen and C. T. Rhodes

Jennifer 5.

Containment in the Pharmaceutical Industry, edited by James Wood Good ~anufacturin~ PracticesforPharmaceuticals:FifthEdition,Revised and Expanded, Sidney H. W~//ig Advanced Pharmaceutical Solids, Jens T, Carstensen

This Page Intentionally Left Blank

nde

edited by

MARCEL

MARCELDEKKER, INC. D E K K E R

NEWYORK BASEL

This book is printed on acid-free paper. Marcel Dekker, Inc. Avenue, New York, NY 10016 tel: 212-696-9000; fax: 212-685-4540 Marcel Dekker AG Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland tel: 41-61-261-8482; fax: 41-61-261-8896 http: l/www.dekker.com Thepublisheroffersdiscountsonthisbookwhenorderedinbulkquantities.Formore info~ation,write to SpecialSales/ProfessionalMarketing at theheadquartersaddress above.

Neither this book nor any part may be reproduced t~a~smitted or in anyform or by any means, electronic or mechanical, including photocopying, micro~lming,and recording, or by any information storage and retrieval system, without permission in writing from the publisher. Current printing (last digit): l 0 9 8 7 6 5 4 3 2 1

The first and second editions

of Drug Stability were authored by Jens T. ~ a r s t e n s e n ~

The considerable success achievedby the first and second editions of Drug ~tabiZity has led to the need for a third edition. Although this edition is firmly based on the foundations laid by JTC, it has undergone a radical change in that additional chapter authors have contributed to the book and CT isnowassisting in the editorial duties. In the period of almost five years that has elapsed since the publication of the second edition, there have indeed been significant developments in t technology, and regulatory aspects of pha~aceuticalstability testing. innovation has not occurred at a uniform rate in all the disciplines o to our field. There have been no fundamental changes in the equations that govern chemical reaction; our understanding of the mechanisms of oxidation has ~ a t u r e d but the underlying theory is still basically unaltered. Important new technologies lly for macromolecules) have been introduced, for stability testing ( however, and have infl the approaches used by pharmaceutical scientists in laboratory stability studies. It is probably in the regulatory arena that the most substantial developments ha curred. The several ICH Guidelines on stability topics and the subsequent document have exerted, and will continue to exert, considerable influence on approaches used by pharmaceutical scientists in stability work. In particular, all the effects of the lengthy F are still not clear. hus, in making our plans for this third edition, we realized that although the discussion of sometopics was in relatively little need of modification from the second edition, others needed radical modification and additional coverage. Further, we felt it appropriate to introduce some topics not previously discussedin Drug ~tabiZityto hope that the third edition not only continues the tradition established in the first and second editions, but also blends in new topics and opinions so that iii

reface

this book can continue to be regarded as a reliable text covering many, if not all, subjects that relate to drug stability. Jens T.Carstensen C. 7". Rhodes

Preface Contributors

1. Introductory Overview C. T. Rhodes

iii vii

2. Solution Kinetics Jens T. Curstensen

3. Kinetic pH Profiles Jens T. Carstensen

Oxidation in Solution Jens I: Carstensen

113

5. Catalysis, Complexation, and Photolysis Jens T. Carstensen

133

6. Solid State Stability Jens I: Carstensen

145

7. Interactions of Moisture with Solids Jens T. Carstensen

191

8. Physical Characteristics of Solids Jens T. Carstensen

209 V

onte~t

9. Prefor~ulation Jens T. Carstensen

237

10. Physical Testing Jens T, Carstens~n

261

11. Development and Validation of HPLC Stability-~ndicati~g Assays Donald D. Hong and M u ~ t a ~ ~ h a h

329

12. Stabilit~Testing of Clinical Trial W o ~ g a n gG r i ~ ~

385

13. A Rational Approach to Stability Testing and Analytical Development for NCE, Drug Substance, and Drug Products: Marketed Product Stability Testing W o ~ g a n gG r i ~ ~ 14. Packaging, Package Evaluation, Stability, and Shelf-Life D. A. Dean

415 483

15. ~ndustrialStability Testing in the United States and ~ o ~ p u t e r i z a t i o n 515 of Stabililty Data Shri C, Val~ani 16. Stability of Polypeptides and Proteins Mary D. DiBiase and Mary K, Kottke

553

17. Regulatory Aspects of Stability Testing in Europe Brian R. M a t t h ~ w s

579

18.

egulatory and Scientific Aspects of Stability Testing: Present and Possible Future Trends C. T, Rhodes

619

Appendix: Guidance for Industry: Stability Testing of Drug Substances and Drug Products (Draft)

63 7

Author Index

751

Subject Index

767

. Madison, Wisconsin ean Consultant, Beeston, Nottingham, England Biogen, Cambridge, Massachusetts

0nd

. Biberach, Germany . Pharmaceutical Consultant, Raleigh, North Carolina . Cubist Pharmaceuticals, Inc., Cambridge, Massachusetts

. Alcon Laboratories (U.K.) Limited, Hemel Hempstead, Hertfordshire, England

. Department of Applied Pharmaceutical Sciences, University of Rhode Island, Kingston, Rhode Island . Trigen Laboratories, Salisbury, Maryland

. Pharmaceutical Consultant, Kalamazoo, Michigan

vii

This Page Intentionally Left Blank

University of Rhode Isl~nd,Kingston, Rhode Island

1. Stability Isan Essential

uality Attribute for Drug Products

2. Potential Adverse Effects of Instability in ~ h a ~ a c e u t i cProducts al 2.1.Lossofactive 2.2. Increase in concentration of active 2.3. Alteration in bioavailability 2.4. Loss of content uniformity 2.5.Decline of microbiological status 2.6. Loss of pha~aceuticalelegance and patient acceptability 2.7. Formation of toxic degradation products 2.8. Loss of package integrity 2.9. Reduction of label quality 2.10. Modification of any factor of functional relevance

2 3 3 5

5 5 5 6

7 7 7 7

3. The Gamut of Stability Concerns 3.1.Bulk drug substance and excipients 3.2. Research and development formulations 3.3.Clinical trials materials 3.5. Reformulation, change of manufacturing site, tro~~leshooting, complaints 3.6. Product in the channel of distribution 3.7. Product under the control of the patient 3.8. In vivo stability 4. Reasons for Stability Testing 4.1. Our concern for patients’ welfare 4.2. To protect the reputation of the producer 4.3. Requirements of regulatory agencies

9 10 10 10

11 11 11 11 1

.4. To provide a database that may be ofvalue in the formulation of 11 5.

5.1. Chemical 5.4.

a1 (especially microbiological) imitations of this classification

6. The Essential Elements of a High-Quality, Cost-Effective Stability Program 6.1. Commitment of the organization to quality 6.2. Firm grasp of underlyingscientific theory 6.3. Up-to-date knowledge of all relevant policies of regulatory agencies and applicable pharmacopoeial standards 6.4. Effective com~unicationbetween R&D, production, QC/QA, complaints, and regulatory affairs 6.5. A general understanding of the limitations of the analytical methods used in the stability testing program 6.6. Careful monitor in^ of the stability budget 6.7. Managerial skills to coordinate and optimize the program

12 12 12 12 12 12 12 13 13 l? 13 14 14

7. Conformance Periods, Shelf Lives, and Expiration Dates

8. Some Possible Strategies to Improve Shelf Life 8. l Sampling and analytical

15 15 15 16 16

8.4. Formulation References

“There never was anything by the wit of man so well devized or so sure established which hath not in the continuance of time become corrupted. ..” Thomas Cranmer Everything made by human hands-from the sublime Parthenon to the trivial milkshake-is subject to decay. Phar~aceuticalsare no exception to this general statement. If there is any functionally relevant quality attribute of a drug product that changes with time, evaluation of this change falls within the purview of the pharmaceutical scientists and regulators who quantify drug product stability and shelf life. The rate at which drug products degrade varies dramatically. Some radiopharmaceuticals must be used withina day or so. Other products may, if properly stored and packaged, retain integrity for a decade or more, although in many

jurisdictions the maximum shelf life that a regulatory agency willapprove for a drug product is five years. (This restriction is hardly an onerous one, since even for a product with a five-year shelf life it is probable that over 95% of the product will be sold and used within thirty months of manufacture, providing all involved in the distribution process obeythe first lawof warehousing: FIFO-first in, first out.) Since the evaluation of the stability of a drug product is highly specialized and esoteric in nature, reliance on the patient’s suck-it-and-see organoleptic evaluation is of distinctly limited value. Thus governments in many parts of the world-most importantly in Western Europe, North America, and Japan-have, because of concerns about drug product safety, efficacy, and quality, found it appropriate to require some form of stability testing for drug products. However, it must be recognized that even before governments became active in this area many reputable companies were already giving attention to drug product stability and developing their own in-house approaches. The increasing intervention by regulatory agencies such as the FDA (U.S. Federal Food and Drug Administration) and the HP (Canadian Health Protection Branch) stimulated standard approaches to stability testing in those parts of the world subject to their control (1). More recently, the process of globalization and harmonization has stimulated the development of world-wide standards. (This topic is further considered in Chapter 18.) It is nowwell accepted that stability is an essential property of drug products; thus the assignment of a shelf life is a routine regulatory requirement.

There is a variety of mechanisms by which drug products may degrade, and thus a quite wide range of adverse effects that can occur.

Obviously, lossof drug is ofmajor importance in the stability studies of many pharmaceutical products. Unfortunately, one sometimes gets the impression that some regard this as the only adverse effect of drug product stability, This is, of course, not true, and for some products loss of active isnot the critical variable that determines shelf life. However,it is certainly true that for many products loss of potency is of major importance. In general, we regard any product that contains less than 90% of label claim of drug as being of unacceptable quality. Therefore, for many drug products, determination of the time that elapses before the drug content no longer esceeds90% (when the product is stored in conformance to label instructions) is an essential element in determining shelf life (2). The essence ofthe conventional way ofdetermining shelf lifefrom loss of active is as follows. The potency of product stored at the appropriate temperature (25°C for products to be labelled “Store at Controlled Room Temperature”) is determined as a function of time and the best straight line of potency as a function of time determined by least squares regression analysis (Fig. l). Of course, because of analytical and sampling error there will normally be some scatter of the experimentally determined data points around the mean regression line.Thus in order to have a high comfort level about the shelf life that we will assign to the product, we use conventional

re 1 Leastsquaresregressionanalysis.

statistical methods to calculate the 90% confidence zone ofthe regression line. This means that there is a 90% probability that the true regression line (of label claim potency as a function of time) is within the zone. Thus there is only a 5% chance that the true line is really below the 90% zone. It is apparent thatany shelf life derived from the intersection of the lower 90% confidence bound and the 90% potency value has a 95% confidence level. In other words, there is only a 5% chance that our estimate of the shelf life will be too high. In fact, there is an additional safety margin built into most estimates of shelf life in that the period of time determined as ce is usually signifidescribed above (termed by A. J.Smith the c o n ~ o r ~ a nperiod) cantly greater than the period definedas the shelf life. Suppose that for three separate pharmaceutical products we obtained 95% confidence estimates of the conformance periods of 13.2, 26.1, and 39.4 months. We would probably assign shelf lives of 12,24, and 36 months to the three products. The difference betweenthe confor~ance period and the assigned shelf life provides an estra stability reserve. Those conversant with kinetic theory as outlined in Chapter 2 may wonder if the order of reaction affects the form of graph obtained as shown in Fig. 1. It might be felt that if the degradation process is governed by first-order kinetics, the plot would yield a curve rather than a straight line. In fact, since this type of plot only covers potency valuesfrom 100% down to go%, both first and zero kinetic processes appear to be essentiallylinear. Yes, there are some degradation processes (3)that do produce nonlinear plots, but fortunately these are remarkably rare. Shelf life values are normally assigned to a product rather than a batch. In. order for such a practice to be legitimate, there must be reliable data that shows that, for at least three batches of the product, there is no significant in. the slopes or the intercepts of the types of plot shown in Fig. 1.In those instances, where this level of batch similarity cannot be demonstrated, it might be possible to assign a shelf life based on a worst-case scenario (i.e., the shelf life is based on the worst-case batch). In this situation, we probably need more than three batches so that we can be comfortable that we have indeed identifiedthe worst-case batch.

ost pharmaceutical products are characterized by onlyoneshelflife. However, in somecases a product mayhavetwo. For example, a freeze-dried (lyophilized) protein product may haveone shelf life, say two years,for the product stored in the dry condition and a second shelf life, say two days, for the product when it has been reconstituted with the appropriate vehicle and is readyfor injection.

For some products, loss of vehicle can result in an increase in the concentration of active drug. For example, some lidocaine gels exhibitthis behavior. Perfusion bags sometimes allow solvent to escape and evaporate so that the product within the bag showsan increase in concentration. This behavior israre, but for such situations a horizontal mirror image of Fig. 1 should be used to estimate shelf life. In such a case, the point where the upper 90% confidence bound intersects that 110% potency value will define the conformance period.

~ioavailabilityand bioequivalence of drug products is a subject of great importance to those concerned withdrug product quality (4). Ifthe rate or extent of absorption that characterizes a product changes on storage, then this is, of course, a stability problem. In particular, if dissolution test data shows significant changes with time, there should be concern about possibleclinicallyrelevantmodification of bioavailability or bioequivalence. ~xaminationof lists of products that have been subjectto recalls showsthat a number of products subsequent to release onto the market have been shown to fail relevant dissolution tests. Suchfailure has resulted in a Class I1 or Class TT1 recall.* Case hardening of the surface of tablets or pellicle formation with hard gelatin capsules have been involved in this type of problem (4,5).It is thus most appropriate to include dissolution (or other release tests) in stability evaluations of pharmaceutical products.

Suspensions are the drug delivery systems most likely to show a loss of content uniformity as a function of time. For suchsystems, determination of ease of redispersion or sedimentation volumemay therefore be included in a stability protocol (6).

In the comparatively recent past, say 15 yearsago, it was relatively commonpractice to give attention to the microbiological status of pharmaceutical products only if theyweredesigned for administration by the parenteral or ophthalmic routes. The situation has now changed, and though we do not expect all pharmaceutical * FDA classifies recalls asI (most serious), potentially with serious, possibly with fatal, consequences; I1 (quite serious), may impair therapeutic response; and I11 (least serious), unlikely to have substantial adverse effect on therapy.

products to be sterile (that is, totally devoid of all forms of life both vegetative and sporing), we do have concern about the microbiological status of all drug delivery systems. We have concernabout the extent of the total bioburden, and we also have a specific interest in excluding pathogens. Basically, there are two possible ways in which the microbiological status of a pharmaceutical product can change significantly with time. First, microorganisms present in the product at the time of manufacture may reproduce and thus increase the number of viableorganisms. Thus a product that, whenassayed for total bioburden at the time of manufacture, iswithinlimitsmay,whentested after say 6 months storage, exceed the maximum permitted bioburden. Second, if package integrity iscompromised during distribution or storage, it ispossible that the microbiological status of the product may be adverselyaffected as a result of the ingress of microorganisms. In order to reduce oreliminate the first typeof microbiological problem, attention should be given to the quality of the raw materials and the nature of the manufacturing facility and its operation. Certain raw materials thatare often the source of microorganisms (both pathogenic and nonpathogenic) are of natural origin (e.g., corn starch, lecithin), and thus there should be particularly careful monitoring of the microbiological status of such rawmaterials. In terms of excluding contamination during manufacture, such factors as positive pressure air flow, equipment design, personneltraining, and clear SOPS (standard operating procedures) all have roles to play.

It is not easy to define pharmaceutical elegance. It includes any aspect of the product that might suggest that the product issomehow substandard or variable. For example, some drugs that contain amino functional groups, when made into direct compression tablets that contain spray-dried lactose, may show some slight yellow or brown specklingon the surface of the tablet. The speckling is causedby the interaction of the drug with a minor component in the lactose, which results in the formation of a chromatophore that absorbs strongly in the visual part of the spectrum. Analysis of the speckled tablets might revealno loss of potency or change in dissolution, but of course no reputable manufacturer will market tablets that look as though they are suffering from measles. Pharmaceutical elegance doesnot mean that all drug products are expected to look and taste nice. Indeed, for some patients, particularly those of the older generation, the reverse is sometimes true. Some patients seem to believe that in order to be effective a product should have an unpleasant smell and taste. No, the important point about drug products is that attributes such as appearance, taste, and smell should be reproducible and not show significant batch-to-batch variation. Also of relevance to pharmaceutical elegance is the ease (or lack thereof) of patient use and any changein patient acceptability. For example, if a topical product exhibits a change in play time or skin drag, this may not necessarily directly affect inherent safety or efficacy, but it may well impair the likelihood of the patient accepting the product or using it appropriately. Similarly, apparently trivial matters, such as a label showing some loss of adhesion at the extreme corners, may not

reasonably besaid to modify the essentialsafety and efficacyof the product. However, it may well engender doubts in the patient’s mind about the quality of the product and thus adversely affect patient compliance.

If a drug degrades to a molecularspecies that istoxic, there must be special attention given to the quantity of such a species found in the product during its shelflife. The classicexampleoften quoted in this regardis the formation of epianhydrotetracycline from tetracycline, although the evidence in support of the alleged toxicity in this example may not be overwhelming. However, with protein drugs it is quite likely that even perturbation of molecular structure in a domain well removed from that responsible for therapeutic activity may result in serious toxic potential. Thus it does seem quite likely that this aspect of stability testing maybecome more important in the future. Concerns about potential levelsof toxic degradation products are probably of considerable importance in the presentreluctance by regulatory authorities to approve stability overages for new products.

Change in packageintegrity during storage or distribution can be a stability problem that may require careful monitoring. For example, if a plasticscrew cap loses back-off-torque, the possibility of chemical or microbiological hazard may be significantly increased.Thus when there is reason to believe that such a problem might exist, it is appropriate to use specific package integrity tests in the stability test protocol. cti

The label of a drug product must be regarded as an essential elementof the product. It provides information on identity, use, and safety. Thus if any aspect of the label deteriorates with time, this can be a serious stability problem. For example, if the plasticizer in a plastic bottle migrates into the label and causes the ink to run and thus adversely affects legibility, this is a major problem.

If there is any time-dependent changeof any functionally relevant attribute of a drug product that adversely affectssafety, efficacy, or patient acceptability or ease of use, monitoring such a change will be within the purview of stability evaluation. For example, when some transdermal patches were firstintroduced into use in the United States, a problem of adhesion aging was observed. Freshly manufactured patches showed excellent skin adhesion. However, in some instances products stored at room temperature for a period of weeks or months showed a loss of adhesion. Thus in use the patches had a tendency to fall off the patient’s skin. For products subject to this adhesion aging, it would obviously be important to include quantification of adhesion potential in any protocol designed to evaluate stability.

For the manufacturer of pharmaceutical products, such as compressed tablets or ophthalmic solutions, the first stability type concern will be specifications for the raw materials from which the drug delivery system isfabricated (i.e., drug substance excipients, water) and the extent, if any, to which suchmaterials may degrade in the period between purchase and use. The seriousness of this type of concern has abated to a considerable extent in the many companiesthat have adopted the just-in-time philosophy of drug product ~anufacture.This method requires that the time for which rawmaterials are stored at the facility where the drug delivery system is to be manufactured is kept to a mi~imum-often only a week or less. This procedure means that whereas in the past it was common to have raw materials that had been purchased as long as two years before use, it is now uncommon in many companies to see any raw material that has been held more than six weeks.Although the just-in-time method of pharmaceutical manufacturing was introduced primarily to reduce the amount of capital tied upin raw materials, it has also had the effect of reducing stability problems with raw materials. However, there are still some raw materials such as proteins (e.g., lecithin) for which stability is still a real concern and for which in-house testing by the pharmaceutical manufacturer may well be prudent. In general, pharmacopoeias such as USP have performed well in providing standards for drug substances, and the tests described in the compendia can often be of great value in assuring the quality of drugs to be used in manufacture of drug delivery systems. The situation concerning standards for excipients is less satisfactory. It is, unfortunately, only comparatively recently that pharmaceutical scientists have given substantial attention to standards for excipients that haye national or international ~ l has,~however, x c ~ i become e n t ~ reco~nition.The andb boo^ of ~ ~ a ~ ~ ~ c e u t i c (6) widely used by pharmaceutical scientists (6) and has attained Also of value in providing information about excipients is the ~aceuticaZAdditives (7). This book provides data on over 6,000 materials and can be a useful source of data for possible standards for excipients and might be used as the basis for a vendor's ~ e r t i ~ c aof t e Analysis. One of the problems that may face a manufacturer of a dosage form is that a vendor of an excipient may be unwilling to provide comprehensive data on the material because of the fear that the i ~ o r m a t i o ncould be exploited by a competitor. Fortunately, there is a rather clever regulatory procedure t Pro The manufacturer of the excipient establishes a .This document contains comprehensive data on suc the toxicology, and stability. Its contents are made available to a pharmaceutical manufacturer refers to the particular excipient in a request concerning a formulation that will contain the excipient.

A c o m ~ o npractice in the pharmaceutical industry is to evaluate a number of formulations for such critical attributes as stability. As timegoesby, those

formulations that are shown to be unsatisfactory are rejected, and thus the number of potentially viable f o ~ u l a t i o n sis reduced. Unfortunately, the formulators in charge of research and development (R& projects do not always inform the quality control chemists responsible for stability testing that certain test formulations are no longer potential candidates to be marketed. This results in a waste of time and money. There is no universally accepted procedure for evaluating stability of formulations. Often, however, acceleratedtesting and/ or ~omparisonswith similar products already on the market can be useful. When a formulation has been finally selected, the official guideline should be followed inobtaining appropriate stability data SO that market approval can be obtained.

Chapter 15 gives specificattention to this important matter.The book Drug Products for CZi~icuZTriuZs (8) may also be of value to readers who are especially interested in this topic.

A very su~stantialpart of the effort of a stability testing program is devoted to producing datathat willconvince a regulatory authority that a particular formulation, process, and shelf lifefor a new or reformulated product are acceptable. Also, once a product is marketed, the manufacturer will wish to generate reliable stability data that will provide assurance that the marketed product continues to justify the shelf life that has been assigned. It is these two linked areas that are a major focus of this book. The process of finalizing the formulation and process intended for the marketed product should be completed as early as possible during clinical trials. The FDA has issued specificrequirements about preapproval inspections that itconducts before an rug Application) or similar marketing approval document is accepted (9). The Agency has also issued (9) useful information concerning its requirements regarding scale-up (1 1).

It w o ~ l dbe naive to think that once regulatory approval has been obtained and a product is on the market, the work of a stability group becomes routine and mundane. Usually the stability of the product is subject to intermittent new studies for a variety of reasons. For example, a decision may be made to modify one or more formulation or process variables, so that, depending on the significance of such changes, stability studies of varying complexity may be necessary. Also, if it is decided to change the site at which the product is manufactured, regulatory agencies mayalso require additional data. The new FDA Guidelines (10) to this point and should, of course, be carefully followedby jurisdiction.

Even if no change of formulation, process, or site of manufacture is contemplated, there may be other reasons for additional stability studies. Unfortunately, it is not unknown for a new product that we believe to have been fully validated with respect to all quality attributes (including stability) to exhibit unexpected stability problems. Theseproblems may progressively developin a most insidiousway,affectingall batches or,in some instances, onlysome batches intermittently. In either event, troubleshooting directed at identifying the cause and then taking appropriate remedial action is necessary. Similarly, if complaints from patients, health professionals, or others involve stability problems, it is obviously important that stability group personnel should be involved in the evaluation of the problem and be consulted when it is decided if remedial action is required. in

It is not sufficient to restrict our concerns about drug product stability to the quality of the pristine, freshly manufactured material that we regard with justifiable pride as it waits in our warehouse for distribution after it has been cleared from quarantine by our Q C l Q A (Quality Control/Quality Assurance) department. Of course, it is normal to store some stability samples in our stability storage areas (retained samples). However, the evaluation of samples that have been stored under the utopian conditions in the manufacturer's stability storage areas is of limited value. Samples retained for stability testing are not dropped off the back of a truck; they are not left on a loading dock in the blazing sun; nor are they left in the freezing cold. Thus it is somewhat unrealistic to expect retained stability samples to reflect accurately the stability status range of products that are in the channel of distribution. As is discussed in Chapter 18, there isnowincreased concern about the stability status of products in the channel of distribution.

There is good reason to believe that, in many instances, the conditions under which patients store their drug products is far removed from optimal. At one time some regulatory authorities were considering the possibility of requiring shelf lives that could be guaranteed right through to the time when the patient used the last dose of the product. It isnow probably generally appreciated that this idea is not feasible. It certainly is, however, most appropriate that pharmacists should take time and trouble to counsel patients on the appropriate ways to store drug products.

The final stability concern is the degradation of the drug in vivo. In particular, the hydrolysis of drug at the low pH conditions of the stomach can be particularly serious. The traditional answer to this particular problem is to enteric coat a tablet, In the past, the materials used as enteric coats were not always effective. The polymers now available for enteric coating are much more reliable (1 1).

Obviously, our primary reason for stability testing should be our concern for the well-being of the patients who will use our products. Sometimes in the mad rush to comply with other requirements, this important fundamental may be discounted or forgotten. Indeed, sometimes one gains the impression that in some quarters stability is regarded as having little clinical relevance. Certainly, if a product that does not degrade to toxic decomposition products and that is not characterized by a narrow therapeutic ratio is present on the market at only85%of label claim, one would not expec atients to be dropping dead in the streets because of this deficiency instability. wever, this is not to say that stability problems can never have serious clinical consequences. For example, inthe early 1980sa packaging stability problem with nitroglyce tablets unfortunately resulted in some nitroglycerin tablets being available in the idwest with potency values of less than 10% of label claim. Since nitroglycerin is used for the emergency treatment of a most serious cardiac condition, angina, there is unfortunately strong cause for concern that some patients may have died as a result of this stability problem. Even if death is not likely becauseof stability problems witha particular drug product, the inconvenience, discomfort, and cost associated withthe use of product that is subpotent or exhibits an unacceptably wide range of potencies may be a serious problem needing radical remedial response. For example, concerns about possible potency problem with L-thyroxeneproducts were of considerable importance in stimulating the FDA torequire, in August 199’7,that all human L-thyroxene on the market products that were on the U.S. market at that time could only remain August2000,unlessnew regulatory NDAs (New Drug Applications) or A s (Abbreviated New Drug Applications) were approved (12). *

We should all be jealous for the reputation that the stability of our pharmaceutical products~ompoundedor manufactured-enjoys. Thus a most important reason for conducting a stability testing program is to assure ourselves that our products will indeed retain fitness for use with respect to all functionally relevant attributes for as long as they are on the market. ul

In many parts of the world, there are legal requirementsthat certain types of stability tests, as required regulatory agencies, must be performed (13).Obviously, the law must be obeyed. wever, it is wrong to abdicate from all scientific judgement and only conduct those stability tests that a regulatory agency is perceivedas requiring. Indeed, there are occasions when any manufacturer with a true dedication to quality will perform stability tests that are over and above those required by regulation.

ata obtained in the stability evaluation of product X in 1999 may prove to be of valuewhen,in 2003, we start developing product Y. There maybeoccasions,

hodes

although they are probably rare, when it will be worthwhile to continue stability testing on an R&D formulation that we know will never be marketed just because we are interested in the stability of a new excipient that we have included in the formulation.

Chemical degradation (solvolysis, oxidation, etc.) is common and is described in subsequent chapters of this book. Our knowledge of kinetics can be of material assistance in dealing with chemical degradation.

Physical degradation can be caused by a range of factors (e.g., impact, vibration, abrasion, and temperature fluctuations such as freezing, thawing, or shearing). Physical testing is described in Chapter 10. Unfortunately, in many instances, our knowledge of the exact mechanisms involved in physical degradation is incomplete.It is also unfortunate that a number of the physical test methods that could be used in evaluation of physical stability (e.g., tablet friability, tablet impact resistance, suspension redispersibility,or injection syrin~eability) arestill nonofficial and variable. It is noteworthy that it was not until 1997 that official standardized test methods for the quantification of bulk and tap density were introduced into the USP, although such tests have value in helping to evaluate compressibility.

In NorthAmerica, Japan, and Western Europe it is microbiologicalfactors that are most likely to be involved in biological stability problems. However, in some parts of the world rats, roaches, ants, and other nonmicrobiological organisms can be responsible for stability problems.

Useful though the above tripartite classification of degradation mechanisms may be, there is a danger that itsuse mayovercompartmentalize our approach to drug product stability. This can be dangerous. In fact, many stability problems involve more than one mechanism. For example, insufficient antioxidant in a rubber condom may result in oxidation of the device by a chemical mode. However,the effect that may be detected is lossof tensile strength, which is, of course, a physical parameter.

It is essentialthat throughout the organization responsible for the development and ~ t quality. production of pharmaceutical products there be a true c o m m i t ~ e to

Absent this commitment, it is likely that astability program will beregarded simply as a burdensome, nonproductive expense. If such an attitudepervades top or middle management-although such attitudes are rarely expressed directly in writing but rather transmitted by a nod and a wink-it is quite possible that the stability testing group will be starved of essential equipment and personnel. One sometimes visits companies where stability testing is two or three months, or even more, behind schedule. Top managers say theycannot understand how the problem has developed since, as everyone in the company knows,their personal dedication to product qualityissecondonly to their commitment toGod, the flag, and the family. Unfortunately, such individuals sometimes “talk big but spend little.”

It should hardly need to be stated that stability testing of pharmaceuticals requires in-depth education in the scienceof pharmaceutical formulation, evaluation, analysis, and statistics. Unfortunately, there still are companies where personnel with such education are lacking.

Official regulations and standard test methods continue to evolve. Thusit is important that atleast one person in every company be charged withthe responsibility of keeping up-to-date files on data from the FDA, the USP, or such other entities as may be relevant that impinge on anyaspect of the design, execution, or interpretation of stability tests. erusal of Pharmacopoeial Forum ( P F ) ,the journal in which provides trailer-type information about possiblenew or mo or monographs, should be mandatory inall companies for which dards may be of relevance. *

Everyonewithanydegree of responsibility for decisions about a stability program-not just those performing the tests in the laboratory-should have a general understanding of the parameters that characterize the test methods used instability testing (accuracy, precision, sensitivity, reproducibility, transferability, etc.). do not require that everyone be expert at say HPLC (high-performance liquid chromatography) or ELISA (enzyme-linked imunosorbest assay), but we should expect that the decision makers be aware of the salient characteristics of the test methods, the results of which are used in decisions about stability.

It is surprising that some companies have no stability budget. It is even more surprising to find that there are scientists designing stability protocols who select test method A instead of test method B (both of which might be technically satisfactory but of significantly different cost), who have knowledge of or interest in the relative costs of the two tests. It is not easy to devise a mechanism for evaluating a stability budget suchthat we can be quite certain that we have accounted for all monies spenton the program. However, even though the budget that we estimate may be relative, rather than absolute, it still can be of substantial value. .7.

The capstone of a high-quality cost-effective stability program must be managerial skills that nurture and coordinate the personal and professional skills of all involved with the program.

The conformance period of drug product isdefined by the mostvulnerable time-dependent quality attribute. As has already been indicated in Sec. 2.1 of this chapter, loss of potency is,for many products, the critical parameter. In those cases where some other attribute is more vulnerable, it will be that property that defines the conformance period. The same generalapproach as thatshown inFig. 1 should be followed; however, instead of plotting potency as a percentage of label claim on the y-axis,one plots the appropriate critical stability parameter. The conformance period is then determined from the intersection of the lowest (or highest) acceptable value of the parameter and the 95% confidence bound of the regression line. In the rather rare event that there are two stability attributes of about the same criticality, then both should be quantified and the lower conformance period used as the basis for the assignment of the shelf life of the product. As has been previouslyindicated, the shelf life assigned to a product is equal to, or less than, the conformance and is usuallya convenient round number (e.g.,7 days, l month, 1 year, 18 months, or 2, 3, or 5 years). The expiration (or expiry) date placed on the label of any givenbatch indicates the date at which the shelf life ends for the batch. Thus if the product is stored in accordance with label instructions, it is expectedthat the product will retain fitness for use up to that date. With the exception of products that have very short shelf lives, it is conventional in many parts of the world to give only the month and year of the expiration date. It is expected that for such dates, e.g., May '03, the product should remain of acceptable quality until the end of the stated month. When products have a 5-year shelf life,the practice of only givingexpiration dates for the months of January or July seems to be becomingmore common. This practice simplifies stock control, since there are fewer dates to dealwith. This approach is used as follows: Suppose we have a product that has a five-year shelf life, and we manufacture batches of the product in February, April, June, August, and November of 2002. The first three batches would be dated January '07; the last two would be dated July '07.

Obviously, if a product is not stored in accordance with labelinstructions, the expiration date cannot necessarily be relied on.

It is fortunate that many drug substances and products are inherently stable; thus, with little difficulty, we can justify a shelf life of 3 years or more. However, there are drug substances that are very much more liable to degradation, and it may require much skill and hard work to develop a product with a shelf life that is commercially acceptable. Since this book is not focused on formulation per se, this section only outlines some of the general approaches that might be considered in efforts to improve shelf life. .l S

~ ~ y t ~ c ~ l

Examination of Fig. 1 reveals that the more scatter that we have on a stability plot the wider the 90% zone of confidence will be. If we were somehow able to obtain experimental points that all fitted exactly on the regression line, then both the upper and lower 90% confidence bounds would also be on the mean regression line; thus our estimate of the conformance period would begivenby the intersection of the regression line and the 90% potency line. Clearly, this would substantially extend the shelflife that we couldlegitimatelyclaim.Ofcourse, it isimpossible to obtain suchperfect data that the 90% confidencezone has no width whatsoever. However, anything that we can do to reduce its thickness will improve our shelf life. There are two main causes for the fact that stability plots, such as that shown in Fig. 1, showscatter, viz., sampling error and analytical error. Anything that we can do to reduce either or both of these errors will improve our shelf life without our having made any change to the formulation or process used for our product. It is not often easy to see how samplingerror could be reduced. Possiblythe use of near-infrared spectroscopy for single-tablet assay (see Chapter 18) of the same known, individual tablets throughout the shelf life testing period, and averaging the data so obtained at each time point, might be a practicable method to reduce error due to content uniformity variation (14). Perhaps reduction of sampling error is one of the incentives that we have in making sure that all samples are tested on time. In terms of analytical error, if we can improve precisionand reproducibilitywe will slimthe 90% confidence envelopeand improve our shelf life.In some cases,it has by the been shownthat the extra cost of a more sophisticated assay may be justified improvement in shelf life that results.

If testing of samplesis continued beyond the point at which degradation has reached the 90% confidence of the label claim value, we move the narrow “waist’, of the 90% confidence zone to later times and thus improve our shelf life. This valid statistical approach wasspecificallymentionedin the 1984 FDA Stability Guidelines.

hodes

Filling ampoules under nitrogen for solutions that aresusceptible to oxidation is one example of a processing method that can improve stability. Of course, the most common processvariable that is adjusted to control stability is selectionof the package components and materials, and readers speci~callyinterested in this topic are referred to Dixie

The literature is replete with accounts of proven and potential methods of improving product stability and hence shelf life. All that is provided in this section are some general concepts that can, if appropriate, be explored in more detail. In the past, stability overages,* which allowed a relatively easy method to improve shelf life, were quite common. Indeed, there are many drug products on the market in different parts of the world that contain a stability overage of up to 10% of label claim. However, a number of regulatory agencies, including FDA, are now showing much more reluctance than previously to approve such overages for drug products.? This reluctanceto approve the use of stability overages probably stems from a number of causes. First, there isconcern about the possibleincrease in toxicity that might accompany the use ofa stability overage. Ifa product for which compendia1 potency limits are 90-1 10% is released onto the market at 100% of label claim, then the ma~imumamount of any degradation product that could be present in the product up until the expiration date is 10%. However, if the product is released at 110% of label claim, then it is conceivable that in some instances there could be up to 20% of degradation product. If the degradation product, or part thereof, is toxic, ~ potential Z e ~ hazard to which a patient is use of a stability overage has ~ o ~ the exposed. Second, if stability overages are allowed, then the range of potencies to which a patient may be subjected is increased. For example, suppose that a patient who has a repeat prescription for drug X (which is known to have a relatively low range of acceptable therapeutic blood levels) finishes tablets of lot A101, which has a potency of 90%, and is then supplied withtablets from B103, which has a potency of 1100/0. Then (even neglecting degradation of drug while the tablets are under control of the patient and not considering content uniformity) we can see that the patient may experience a 20% variation in blood levels. In contrast, in the absence of a stability overage, the maximum potencyvariance would only be10%. This substan* Overages are of three types: container, man~acturing,and stability. A Container overage is added to allow for the fact that it is not possible in some cases to remove all the contents from a container. Thus ampoules labeled 1.0 mL are normally filled with 1.1 mL. A ~ a n u f f f c t ~ r ioverage ng is added when it is known that relatively small and reproducible amounts of active are always lost during the manufacturing process although we are using modern equipment and facilities and well-trained staff. A manufacturing overage is, of course, dissipated by the time final product testing is completed. t Vitamin products, which are classified by FDA theas food supplements (unless they are administered by the oral route or supplied under a doctor’s prescription), still have substantial overages-sometimes to up 100% of label claim,

tial potency variation could lead to sub- or supratherapeutic bloodlevels and perhaps the need to retitrate the patient. Third, and perhaps most important, there is a perception in some quarters that use of a stability overage isa cop-out that represents an easy Band-Aidapproach to formulation that is quite unacceptable in modern pharmaceutical technology. It is thought that a more thorough investigation of the problem and a willingness to devote appropriate resources of time, personnel, and money might well allow the problem to be solved by other, more conventional, formulation approaches that do not require a stability overage. Formulation approaches to reduce the problem of hydrolysis of drugs in solution have generally been of rather limited success. Recently, complexationof drugs with cyclodextrins has attracted considerable interest (15). Such complexes may show improved resistance to hydrolysis, faster dissolution., and better bioavailability. Of course,sincemosthydrolysis reactions are catalyzed by hydronium and hydroxyl ions, pH control might appear to have great value as a formulation approach to reducing hydrolysis. In practice, however, this approach has had rather limited success. For drugs liable to hydrolysis that are formulated into tablets, the use of a coating may be of value in improving stability. In contract to hydrolysis, degradation by oxidation can often be successfully controlled by formulation approaches. There is a range of chelating agents and both oil- and water-soluble antioxidants that areused in products in various parts of the world. When a product contains an antioxidant, it is normal to monitor the amount (or concentration) of antioxidant as part of stability studies. In theory., it would be acceptable if all the antioxidant were used by the end of the shelf life period. In practice, most of us would feel rather uncomfortable ifwe did not have, say, 25% remaining at the end of the shelf life. Antimicrobial preservatives, suchas sodium benzoate, are commonly added to many pharmaceutical products. The amount (or concentration) of such components should be monitored during stability studies. Although chemical assay for antimicrobial preservatives may be acceptable at most timepoints, the testing performed at the last time point should be by a microbiological challenge test, such as that specified in the USP. Perhaps the area where formulation approaches are particularly important in controlling stability problems is the field ofprotein drugs, an areaof ever-increasing importance. Dr. Kottke and Dr. DiBiase give this topic specific attention in their chapter in this book.

1. FDA Guidance for Industry. Draft Stability Testing of Drug Substances and Drug Products, 1998. 2.B. ~ommanaboyina,C. T. Rhodes. Drug Devel. Indus. Pham. 25, in press. (1999). 2. 3. C. M. Won, Pharm. Res. 9:131-137,1992. 4. J. T. Carstensen, C. T. Rhodes. Drug Devel. Indus. Pharm. 19:2709-2714,1993, 5, S. E. Tabibi, C. T.Rhodes. In: Modern Pharmaceutics. 3 ed.G. S. Banker, C. T. Rhodes, eds. New York: Marcel Dekker, 1995. 6. The Handbook of Pharmaceutical Excipients, 2nded. American P ~ a r ~ a c e u t i cAssocial ation and the Royal Pharmaceutical Association of Great Britain, 1996,

7 . M. Ash, I. Ash. Handbook of Pharmaceutical Additives. Gower, Croft Road, England, 1995. 8. Donald C. Monkhouse and C. T. Rhodes, eds, Drug Products for Clinical Trials. New York: Marcel Dekker, 1998. 9. Martin D. Hynes, 111. ed. Preparing for FDA Pre-Approval Inspections. New York: Marcel Dekker, 1998. 10. FDA.. SUPAC Guidelines, 1995. 11. C. T. Rhodes, S. C. Porter. Drug Devel. Indus. Pharm. 24:1139-1153, 1998. 12. C. T. Rhodes. Clin. Res. Drug Reg. Affairs 15, 180-185 1998. 12. ICH Harmonized Tripartite Guideline for Stability Testing of New Drug-Substances and Products, 23 September, 1994 (ICH QIA). 14. K. M, Morisseau, C. T. Rhodes. Pharm. Tech (Tabletting Yearbook) 6-12, 1997. 15. M. D. Dhanaraju, K. Senthil Kumaran, T, Baskaran, M. Sree Rama Moorthy. Drug Develop. Indus. Pharm. 24583-587, 1998.

~ ~ d i s o Wisconsin n,

1. The Order of a Reaction

2. The Zero-Order Reaction

3. First-Order Reactions

4. First-Order Reactions with More Than One End Product 4.1. Consecutive reactions of the first order 4.2. Parallel reactions

27 27 29

5. Equilibria 5.1. Steady-state situations

30 33

6. Pseudo-Zero-Order Reactions

7. The Arrhenius Equation 7.1.Cyclic testing 7.2. ~ o n i s o t h e ~kinetics al 7.3. Kinetic mean temperature 7.4.Eyring plots

37 39 41 42 45

8. Second-Order Reactions 8.1, Equal initial concentration. Reciprocal plots 8.2. Pseudo-first-order reactions 9. Complicated HydrolysisSchemes

46 52 52 53

References

lability is not synonymous with chemical kinetics, yet most of the rate-limiting phenomena are either associated with chemical reactions or aredescribable by some equation system that bears a resemblance to those encountered in chemical kinetics. It is, therefore, of importance to lay the proper kinetic foundation before discussing the actual phenomena encountered in dosage forms. These fundamental principles are most conveniently described by solution kinetics. The simpler a system is, the easier it is to make it reproducible, and it is therefore not surprising that the largest number of pharmaceutical publications on the subject of kinetics deal with solution systems. Furthermore, the more dilute a system is, the more it will adhere to ideal laws, and hence the largest number of publications to be found deal with dilute systems. There are obviously pharmaceutical dosage forms that are solutions, viz. oral, parenteral, nasal, ophthalmic, and otic solutions. Of these, it is only the parenteral and ophthalmic solutions that are chemically fairly simple, i.e. contain only a few number of components. These are systems that would behave similarly to the patterns described in, for instance, the chemical literature. In oral solutions, there are many ingredients (sweeteners, solubilizers, etc.), so that, here, one would expect definite vehicle effects and interaction possibilities. The Stability Guidelines make certain requirements on basic stability that are best elucidated (or only elucidated) through solution kinetics: First of all it is necessary to develop a stability-indicating assay. This is defined in lines 111 of the 1987 Guidelines as “~uantitativeanalytical methods that arebased on the ~haracteristic structural, chemical, or biological properties of each active ingredient of a drug product and that will distinguish each active ingredient from its de~radationprodso that the active ingredient content can be accurately measured.” The 1993 Guidelines state, Analytical test procedures should be fully validated and the assays should be stability-indicating. The need or the extent of replication will depend on the results of validation studies (194-196). The focus may instead be on assuring the specificity of the assay ... of identified degradants as indicators of the extent of degradation via particular mechanisms (386-389).

This means that the assay must be capable of detecting quantitatively the amount of parent drug present, and identify, and to some degree quantitate, the decomposition products. Lines 265-277 of the 1987 Guidelinesstate, dation products are detected, the following information about them should be submitted when available: (a) Identity and chemical structure, (b) cross-reference to any available information about biological effect and significance at the concentrations likely to be encountered, (c) procedure for isolation and purifi~atio~, (d) mecha~ismof formation, including order of reaction .., (e)physical and chemical properties, (Q specifications and directions for testing for their presence at the levels or concentrations expected to be present.” Lines 141-144 further state that “the stabi1it~-indicating~ethodologyshould be

validated by the manufacturer (and the accuracy and precision established) and described in sufficient detail to permit validation by FDA laboratories.” In developing stability-indicating assay methodology, it is customary to deliberately decompose the drug in solution, so asto challenge the assay and insure its capability of separating the parent drug from decomposition products. It is obvious, also, thatit isdesired to establish the kinetic order of the decomposition.

The order of a reaction will be defined below, but in essence it determines how the degradation data aretreated. at it is important toestablish the order of a reaction Guidelinesspecifically state, isevidentin that the 1993 I The natureof any degradation relationship will determine the needfor transformation of the data for linear regression analysis. Usually the relationship can be represented by a linear, quadratic orcubic function on an arithmetic or logarithmicscale. Statistical methods should be employed to test the goodness of fit of the data on all batches and combinedbatches(where appropriate) to theassumed degradation line or curve (138-143).

They also state (in respect to mass balance), This concept is a useful scientific guide for evaluation data but it is not achievable in all circumstances. The focus may instead be on assuringthe routesof degradation, andthe use, if necessary, of identifieddegradants as indicators of the extent of degradation via particular mechanisms (385-389).

Although the presentation modes outlined in this quotation are not (at least not in the case of the quadratic or cubic functions) of scientific bent, it is obvious that efforts must be made, before the formal stability program is started, to establish the order of the reaction. ~stablishingthe order is, furthermore, of financial importance, because the later) depends, to establishing of expiration periods (whichwillbediscussed some degree, on the investigator’s capability of extrapolating the concentration of drug beyond the last time point of testing. The 1993 ICH Guidelines further state, Limited extrapolation of the real time data beyond the observed range to extend expiration dating at approval time, particularly where the accelerated data supports this, may beundertaken. However, this assumes that the samedegradation relationship will continue to apply beyond the observed data andhence the use extrapolation of must be justified in each application in terms of what is known about the mechanism of degradation, the goodness of fit of any mathematical model, batch size, existence of supportive data, etc. (149-155).

The longest possible expiration period is, of course, economically desirable, and many of the efforts of the stability programs of pharmaceutical companies are geared towards lengthening this period. As for definition of the order or a reaction, if (2.1)

then the reaction rate is given by

where C is the concentration of the species being studied, brackets denote concentrations of A and B, and k denotes a rate constant, then the reaction is said to be of the order n + m . The rate constant, in this writing, will most often carry the subscript denoting the order of the reaction whenever reaction orders are discussed and being distinguished, ( A notable exception is the notation in the section dealing with pH profiles). The most important orders of interest in the pharmaceutical sciences are integral orders, i.e. those in which the sum of n and m is 0, 1, or 2. (Orders of higher than two are rare.) As alluded to above, knowledge of the order of a reaction is of great importance in stability determination of drug substances, in particular in solution. The problem is frequently to judge whether the concentration-time profiles are linear (zero order) or curved (first or other order). When large amounts of data are at hand (e.g., at different temperature, where the order does not depend on temperature), then a data-consolidation technique described by Carstensen and Franchini (1994), Carstensen (1997), and Franchini and Carstensen (1994, 1999). The technique has later been usedby Shalaev et al. (l 997) in the study of solid-state methyl transfer reactions. For instance, the data in Fig. 1, when plotted linearly, give fairly good plots, but since there are only a few points it is difficultto say with reasonable certainty whether the data are, indeed, linear or curve-linear. Afr~ctionallife is the length of time it takes for a product or drug substance to decrease to the level indicated by the fraction: the half-life, t50, of a substance is the length of time it takes to decrease the content of active compound to 50% of its value. If, in Fig. 1, a given fractional life (e.g., t90 as shown in the figure) is read

Time (Months) Example of sparse data atthree temperatures, such as is often encountered in early development of drug products.

able 2.1 Example of Reduced Data Treatment in Kinetics

___~

______

Temperature Time (Months)

(GO)

1 (3.5) 25°C

0 3 6 9 12 0 1 2 3 0 0.5 l

3'7°C (1.7)

1(0.95) 45°C

Fraction retained

Reduced time 1.ooo 0.905 0,819 0.741 0.670 1.ooo 0.935 0.875 0.800 1.ooo 0,942 0.887

.ooo 0.905 0.819 0.741 0,670 1.ooo 0.935 0.875 0.800

.ooo

0.942 0.887

off the graphs, then the data may be consolidated. The

t90

values shown are 45°C:

0.95; 35°C: 1.7; and 25°C: 3.5.The data in the first column in Table 1 are then

reduced to the fourth column, and this is used as abscissa and the fraction retained (regardless of temperature) is used as ordinate in column 3 , and the data are shown in Fig. 2. The method allows better extrapolation tolerances, since the number of points is larger than for the individual temperatures. If, as in the unstable shown in Fig. 1, the t90 value is 3.5 months, then extrapolation could be made to 2 4 / 3 3 =6.7 half-lives, and the estimated potency after 24 months could beestimated with better precision than if only the 25°C data had been used.

y = 1.0012 0.11875~+ 4.9675e-3xA2 RA2 = 0.999(Curve-Linear) y = 0.99595 - 0.10421~ RA2 = 0.996(Linear)

Reduced Time (t/t90)

Data from Table 1 plotted by reduced time treatment.

The fractional life should be read off the graph (Fig. 1) since at thatpoint the reaction order is not known. Oncethe data are plotted, it is possible to execute curve fitting and estimate the best fit. In the case cited, the number of points is probably still too small to make a decision, but the indication is that the data arefirst order. Including data from even higher temperatures will help in this respect, but it is necessary that the order of reaction not change at the higher temperatures. It has been mentioned that above 85% it is difficult to distinguish between different reaction orders. Li et al. (1998) report on apparent first-order plots for oxidation of a 4-[2-(2-amino-4-oxo-4,6,7,8-tetrahydro-3~-pyrimidino[5,4-6][1,4] thi~zin-6-yl)-(S)-ethyl-2,5-thenoyl-~-~lutamic acid. It is to be noted (from their Fig. 6) that there is definite downward curvature in the plots, and that they probably are S-shaped, as is discussedfurther in the chapter on oxidation. a method of fractional times will help in deciding on which orders are plausible, or if a certain order of reaction can be ruled out. It should, finally, be mentioned that ~ a l k k i - ~ a m and e Valkeile (1988) have described a method for transforming regression curves to the determination of reaction order of given situations in stability studies, using the ox-Cox technique and the Link function transformations.

There are notmany truly zero-order reactions in the pha~aceuticalfield. It will be shown at a later point that there are several types of reactions that will appear to be zero order, i.e., are pseudo-zero order. The equation for zero-order reactions is dC -ko

where C is concentration, t is time, and ko is the zero order rate constant. It is seen that the unit of k is concentration units per time unit, e.g. molar per second. The integrated form of Eq. (2.3) is

where a is the fraction remaining at time t,. A quantity often utilized is the half-time, t1/2, which is given by

It is noted that this is dependent on the initial concentration. Zero-order data may be graphed on plain Cartesian graph paper, using concentration as ordinate and time as abscissa. An example (Higuchi and Rheinstein, 1959) is shown in Fig. 1.

lutio

tic

In this case Eq. (2.2) takes the form

which integrates to l n [ ~ ]= -kit t is noted that the a-fractional life is given by

The most common of the a-lives isthe half-life and the t90 (i.e., the point where 90% of the original concentration is left), which adhere to Eq. (2.9) by the equations k l t p = -0.693 klt0.9

= -0.105

ig. 3 and Table 2 show an example of a straight first-order reaction. In stability situations it is required to monitor both the disappearance of the drug and the appearance of decomposition product(s). In most cases there is more than one deco~positionproduct (and simple cases of this will In the simplest case there is only one decomposition product. this, e.g., aspirin in simple systems(Carstensen et al., 1985;Carst 1988a, 1988b) decomposes,by a pseudo-first-order reaction, in a simple fashion, i.e. to salicylic acid and acetic acid. In such cases, if the assay of the decomposition product is fairly good, the decomposition can be monitored best by monitoring 50

250

Hours

500

750

~ e c o ~ p o s i t i oofnvitamin A acetate (to anhydrovitamin A). The least squares fits are (with 5% water) y =0.48 0.015x; (without water) y =0.75 0 . 0 5 ~(Figure . constructed from data published by Higuchi and Rheinstein, 1959.)

~rstensen

able 2.2

Decomposition of Decarboxymoxalactam

Time (Min)

YORetained

In[% Retained]

% Decomposed

100 78 50 38 27 17

4.61 4.36 3.91 3.64 3.30 2.83

0 32 50 62 73 83

10 20 30 40 50

Source: Reconstructed from data published by

Hashimoto et al. (1984).

the appearance rate of the decomposition product, which should follow the reaction (2.12) exp(-kl t)] [B] = Ao[1 An example of this is shown in Table 2. It should be noted that whenever this approach is taken, it is mandatory still to monitor the content of parent drug, because mass balance should persist throughout the reaction period. (If the molar quantities do not sum up to A0 (within experimental error), then either the reaction is not simple A -+ B, or the analytical procedure fails in aged samples). The easiest way of plotting the data in Table 2 is obviously to subtract each [B] figure from A0 and plot is as [A]. One might then argue that one might simply plot the experimental valueof [A]. But for fairly stable systems, the values of[A] may not differ (decrease) much and may be masked by experimental error. The percentage changein [B], however, issubstantial, as seen in the table, and plotting becomes more meaningful (see Fig. 4). There are many reported instances of first-order reactions in solution. For instance, Jordan (1998) has shown that timolol and propanol01decompose by straight first-order kinetics at pH 7.4. Aso et al. (1997) have shown that aqueous solutions of cephalothin decompose by first-order kinetics and that they follow

Time ~~inutes)

Plot of first-order data from Table 1. (Graph constructed from data by ~ a s h i m o t et o al.,1984.)

an Arrhenius equation. Hammad and Miiller (1998) have found clonazepam to degrade first-order in phosphate buffers at pH 7.4 and to adhere to an Arrhenius equation. Heat conduction microcalorimetry has been usedas a method to evaluate stability and excipient stability by a series of researchers Angerg et al., (1988, 1990, 1993). Hansen et al. (1989), and Wilson et al. (1995) have described the general method and results interpretation. Oliyai and Lindenbaum (1991)have studied the decomposition of ampicillin in solution by means of microcalorimetry.

The considerations above have assumedthat the scheme is simplya reaction of type A -+B, but often there is more than one decomposition product. @ ~ c t i oof~The s Firs

The 1993 ICH Guidelines state that mass balance (or material balance) is The processof adding together the assay valueand levels of degradation productsto see how closely theseadd up to l00 per cent of the initial value, with dueconsideration of the margin of analytical precision (382-384).

It is possiblethat the primary decomposition product itself isnot stable, and in such cases the reaction scheme is (2.13) In other words, there will be more than one decompositionproduct. If all the products can be identified and quantitated, then it follows that the number of moles of A, B, and C should always add up to the initial numberof moles of A. It is noted that it is the number of moles that must add up. Addition on a weight basis would be futile if there is a substantial difference between the molecular weights of the drug and the products. The guidelines recognize that it can be difficult, at times, to ascertain mass balance, partly due to analytical precision. More often it is “unknowns” that cause the problem. If C were not identified, for instance, and was detected as a peakin a HPLC chromatogram, then its “content” is often stated as the area under the peak, using the drug as the unit of measure. But if, for instance, a U V detector is used, and C is lackingthe amount of chromophores that A possesses, then the area under the C peak may grossly underestimate the amount of C. An example of this is chlorbenzodiazepine, which hydrolyzes to the lactam form, and then further to the benzophenone (Carstensen et al., 1971). In fact in this reaction, for some ofthe benzodiazepines, C can progress further with the formation of the carbostyril and the acridone derivative, and some of the steps are associated with equilibrium conditions. The rate equations governing scheme (2.13) are (2.14)

= -k2[B]

+kl[A]

(2.15)

and

d[C1 -k2[l3] dt

(2.16)

These simultaneous differential equations maybesolved means and yield the following results:

by conventional (2.17) (2.18)

and (2.19) It is noted that the above expressions refer to molar quantities. An example of consecutive reactions is shown in Table 3 and Fig. 5. The table shows C = 100 -[A]-[B]. It often happens that one of the decomposition products is difficult to assay for, and in such a case, it may be obtained by difference, provided that mass balance is checked occasionally, e.g., in the early stages and at the end. Of course, there are reactions that have a multitude of end products, and in such cases it is conventional to assume that if e.g. a HPLC peak is less than 0.5% then it is considered negligible. This may be dangerous, because (especially ifit is a constant wavelength peak), the actual molar content of the product(s) in the peak may be more than 0.5% (in which case mass balance would be lost). To ascertain that the unidenti~edproducts are not toxic (and since they are unidenti~ed,specific toxicitycannot be checked),it is conve~tional,as well, in such cases, to degrade a sample considerably and check its toxicity. It is worth~hile,

le Time (hours) 0 0.25 0.5 0.75 1 1.5 2 3 4

P~otolysisof Cefotaxime Cefotaxime A (%moles)

Anti-isomer, B (YOmoles)

C by differencea

100 82 70 55 43 28 20 10

0 10 15 18 19 18 15 10 5

0 8 15 27 38 54 65 80 90

Column 3 not reported by Lerner et al. (1988). C may be more than one product, Source: Constructed from data published by Lerner et al. (1988).

;ij 0

Hours The A -3B part of cefatoxime photolysis.Ctotalhas been obtained as 100-[A]-[B], but this doesnot accountfor the possibilityof other reactions of C. (Graph constructed from data by Lerner et al., 1988.)

however, also to check the toxicity at intermediate points, because C might be toxic, but degrade into nontoxic products, and the toxicity of a partly degraded sample might be worse than that of a fully degraded sample). eaction is rather common; for instance, it has been reported There continues to be, in present literature, reports of this type of reaction; for instance, Archontalci et al. (1998) reported on the decomposition of nordaze showed typicalA.-B-C plots with the A degradation being first order, the having a maximum, and the C profile havingthe typical upswing. n the decomposition of theo-m-GLA. and found it to be biexponential. and Bundgaard (1984) and Beal et al. (1993, 1997) reported that the hydrolyses of 3-acetyl- and 3-propionyl-5-FU were biexponential and found that an initial equilibrium of 3-acyl-5FU with 02-acyl-5FU, which then hydrolyzed to 5-FU, explained this.

If A can decompose into two species, B and C, then the reactions may be represented by: A -+ B

kl) (rate constant

(2.20)

A -+ C

(rate constant kz)

(2.21)

and

The rate equation is

Parallel Reactions (5-azacytosine decomposition) Time (hours)

5-azouracil

5-azacycytosine (x lo4 Molar)

(x lo4 Molar)

1.65 1.4 1.18 1.oo 0.85 0.72

0 0.13 0.20 0.25 0.27 0.29

0 0.5 1 1.5 2 2.5

Source: Table constructed from

Nonchromophoric compounds (x lo4 M ) 0

0.3 0.5 0.6 0.67 0.72

data by Notari and de Young (1975).

which integrates to (2.23) At any given time the (molar) ratio of formation of B and C is given by (2.24)

An example of this is shown in Table 4 and Fig. 6. Other examples are those of Visconti et al. (1984), who havestudied the degradation profile of cadralazine in aqueous solution. The reaction consists of four parallel reactions. Fabre et al. (1984) have shown that 3-aceto~ymethylcephalosporin, cefotaxime sodium salt, in aqueous solution, decomposes by the scheme Cefotaxirne -(k2)”+lactone

-(k++products

Cefotaxime -(k2)”+products i.e., a combination of a parallel and a consecutive reaction. IA

Frequently a reaction will proceed and level off. In such cases there is often an equili~rium:

A-B

(2.25)

with an equilibrium constant, K, given by (2.26) enoting the forward rate constant k, and the backwards rate constant k, it follows the amount going to the right in that when e~uilibriumhas been achieved(at t = m), the reaction must equal the amount going to the left, i.e.

Hours

ig. 6 Example of parallel reactions: decomposition of azacytosine. (Graph constructed from data published by Notari and deYoung, 1975.)

or (2.28) Denoting by A, the infinity concentration of A (and hence by A0 concentration of B), Eq. (2.2'7) may be written

A, the infinity (2.29)

which may be rewritten (2.30) The rate equation for Eq. (2.25) is

(2.31)

which integrates to

ln[A -A,] = [k> k,]t

+ In[& -A,]

(2.32)

or (2.33) The work regarding the hydrolysis of hydrocortisone butyrate by Yip et al. (1983) is an example of this type of decomposition combined with an A -+ B -j.C

Decomposition of Progabide in pH 1.75 Buffer Time (min)

Concentration C (Molar)

0 7 13 19 25 31

ln[C-0.0055] -5.666 -6.166 -6.502 -6,812 -7.386 -7.601

0.00896 0.0076 0.00700 0,00660 0.006 12 0.00600

Source: Table constructed from data published by Farraj et al. (1988)

reaction, with the equilibrium occurring between A and B. Ghebre-Sellassie et al. (1984) have described the epimeri~ationof benzylpenicilloic acidin alkaline media and shown it to be an equilibrium between5R,6R-benzylpenicilloicacidwith pena~aldicacid (enamine). Table 5 and Fig. 7 show data by Farraj et al. (1988) showinga leveling effect. If the data in Table 4 represent a simple equilibrium, then the equilibrium level could be obtained by iteration, by assuming different values for the equilibriu~level and choosing the one giving best linearity in the form In[Coo- = --kt

+ h[C,

-CO]

(2.34)

In this case C = 0.0055 and the data areplotted in this fashion in Fig. 8, but it should be underscored that they have simply been used as an example. The alkaline hydrolysis of chlorambucil (Owen and Stewart, 1979) isanother example of an equilibrium situation. Beal et al. (1993) tested the hydrolysis of 3-acetyl~5-fluorouraciland showed equilibriu~kinetics, as did Prankerd et al. (1992) in the case of rifampicin. 0.009

0.008

0.007

0.006

.""""""""~""~"

0.005

Time (rnin)

7 Decomposition of progabide in pH 1.75buffer.(Graph. reported by Farraj et al., 1988,)

constructedfromdata

Time (min) Data from Fig. 6 treated using a niveau level of C = 0.0055. Least squares fit is

y = -0.764 =0.0127~(R== 0.97). (Graph constructed from data reported by Farraj et al.,

1988.)

.l.

If a situation occurs where A -+ -+C and the latter is fast, the kinetics can be simplified by assuming that [B] is "at steady state" throughout the time course. This, obviously cannot be true at the onset. The equations governing this situation are dEA1--kl[A] dt

(2.35)

d[:Bl=kl[A]-k2[le] = 0

(2.36)

= k2[B]

(237)

where the steady state has been imposedby setting the expression in Eq. (2.36) equal to zero. Hence [A] = Age-k1t

(2.38)

and since it follows from Eq. (2.36) that (2.39) then Eqs. (2.38) and (2.39) inserted into Eq. (2.37) give

which integrates to (2.41) i.e., the reaction occurs as if B were not in the picture at all. In general, if there is sucha fast step in the first step of a complex reaction, it is not incorrect to consider it an A-B-C reaction. arguments can be made in the case of A -+ -+C -+D reaction was much more rapid than the others, and i uch a case it would be justified to think of this as an A -+C -+ reaction. To be more exacting, a steady-state approach would probably be better. It should be pointed out however, that the steady-state approach is a fundamental approximation, and if it is used, then the reasonableness of the approximation should alwa S be checked. ichaelis- ent ten type Th y-state approach is often used, particularly ,as an example, let usconsider the situat often occurring, that kinetics. many low level decomposition products are encountered. There are different regulatory views on this, one being that no more than 1% of a product may be formed for it to be considered a minor decomposition product. The situation is hazy, at best, at all times, because often the compounds are unknown. In such cases the "amount" of the decomposition product in the small peak is estimated by the ratio of its area to thatof the main peak. But if the decomposition product has a different A, then this estimate is incorrect, and this is likely to occur if, for instance, in an HPLC setup a single-wavelength UV detector is used. A well documented and elucidated example isthe case reported by Vilanova et al. (1994),whoshowed in alkaline hydrolysis of cefotaxime the presence of deacetylcefotaxime, the 7-epimer of cefotaxime,the 7-epimer of deacetylcefotaxime, the exocyclic methylenecompound, and examine compounds. With such an array of decomposition products, it is important to establish the major products, and treat, in approximation, the decomposition in this light. I the simplest case, cefotaxime shows an A-B--C and A-D-E reaction, with two curves and three C curves. Another case that serves as such an example isthe case of relaxin oxidation by hydrogen peroxide reported by Nguyen etal. (l 993) shownin Fig. 9 where there are two intermediates (B and C) showing maxima and a final product, D, showing the monotonically increasing pattern.

Whenonlysmall amounts of decomposition occur, it isdifficult to distinguish between zero-and first-order reactions. This is because for small valuesof x (-c0.15) ln[l -x] z -x

(2.42)

where x is the fraction decomposed. If the initial amount of drug substance is Ao, then the fraction decomposed is A0

-A -1"- A

-x

A0 A0

(2.43)

100

2 0 0 300 Time, min

400

500

Decomposition of an A "+B; A "+C ; B "+C "+D reaction. (Graph constructed from data published by Nguyen et al., 1993.)

or A = 1 -x A0

(2.44)

A first order reaction would require that

In-

= -kt

(2.45)

but this may, via Eq. (2.37), be written ln[l -x] M --x= --kt

(2.46)

or (2.47) which may be written

A = A0 -Aokt

(2.48)

i.e. a zero-order reaction. Since it actually was a first-order reaction [Eq. (2.45)] such a situation is referred to as a pseudo-zero-order reaction. A set ofdata is shownin Table 6, treated in zero-order fashion in Fig. 1OA and in first-order fashion in Fig. 10 ;and it is seenthat the fits are comparable. The least squares fit data are shown in Table 6. It is noted that different time intervals are used for the different temperatures, and it is one of the tasks, before starting studies at higher temperatures, to establish what the time intervals should be. There is no sense in e.g. testing at 3, 6, and 9 months at 55"C, if all the drug is lost after 3 months' storage.

.6 Assays for anArrhenius Study Potency (“C) 37

Months 25

0 100 100 100 3 99.5 6 93 99 9 98.5 12 99.5 98.5 18 97.5 24 99.0 96.5 36 98.5 48 98.0 1.17 ko %/m0 0.05 0.0017 0.042 0.16 -0.69 -3.17 WkOI -1.77

100 98.5 97 95.5

96.5

3 1.10

100

Months

Data from Table 5 treated by zero-order kinetics.

4.60

1 15°C

5°C 4.50

4.40

100 91 83

Time (months)

Data from Table 5 treated by first-order kinetics.

ate constants are, of course, a function of temperature, and the data shown in Table 5 are graphed in Figs. 10A and 10B. If rapid results are desired for a given product, it is at times a practice to store it at elevated temperatures. The purpose of this is to force sufficiently large degrees of decomposition in a short time, so that they may be assessed with accuracy. The data in Table 6 are artificially precise, and with a bit of assay error, the 25°C data would not show a discernible loss after 6 months. Is it possible to get some idea of what the loss would actually be, and what it would be after 24 months, without having to wait too long? To get an answer to this (an estimate, not a precise answer) is one of the reasons that Arrhenius plotting is carried out for drug products. method is actually quite precise in solution systems. The temperature dependence of a chemical reaction (as long as it is the rate-determining rate constant that is beingtreated) follows the so-called Arrhenius equation given by

(2.49) Ea ln[Z] ln[k] = RT or its antilo~arithmicform, (2.50) y, R is the gas constant, and Tis the absolute temadding 273.15" to the degrees Celcius (Centigrade). 01 Tis used (becausethe numbers then are between 2 and and 0.004 and hence are easier to handle). The slope of a plot according to Eq. (2.42) is still Ea/R, but Eawill now be in k a1 (rather than in cal) per degree per mole. An exampleof this type of treatment of the first-order data in Fig. 1OA is shown in Table 7. A similar table may beconstructed for zero-order treatment, and the graphical presentation is shownin Fig. 1OA. When the rate constants are plotted according to Eq. (2.49), then Fig. 11emerges. Least Squares Fit Parameters from Fig. 10B Temperature "C 15 25 37 45 55

Temperature "K 288.15 298.15 400.15 408.15 418.15

k (mo") 0.00042 0.0014 0.0052 0.01 18 0.031

1000/ T 3.473 3.356 3.224 3.145 3.049

WkI

-7.783

-6.571 -5.259 -4.440 -3.47 1

2 1

3.0

3.1

3.2

3.3

1ooorr

3.4

3.5

Data from Figs. 10 A and I3 treated by the Arrhenius equation.

The Arrhenius equations for the two types of plotting are 1) -lO.O3(1000/T) (2.5 ln[ko] = 31.63 -10.2(1000/T) ln[kl] = 27.631 (2.52)

It is noted that both plotting modes give good results and aboutthe same activation energy. Using a value of R = 1.99 cal/degree-mole gives an activation energy of o n energyfor many reactions. about 20 kCal/mole. This is quite a c o ~ ~ activation If only the data at the three high temperatures had been present, they could have been plotted as in Fig. 11 and extrapolated to 25°C (where 1000/ T= 3.356). Inserting this into Eq. (2.49) gives [kzs]= 27.631 -10.2 x 3.356 = -6.601 or [kzs] = 0.00136 which is close to the value in Table 7. One can now construct a curve, as shown in Fig. 12, which isan extrapolated curve. The data points from Table 6 are shown for comparison. In general, extrapolations are not that good, but in solution systems they frequently approximate the curve that, after time has elapsed, is the actual curve. A case in point is flurogestone decomposition at pH 7.3 reported by abadi et al. (1984). Arrhenius plotting can also be carried out by using t90 data. Since kl t90 = 0.105 it follows that ln[kltg~]= ln[t90]

Ea + ln[kl] = ln[tso] -+ ln[Z] = -0.105 RT

or Ea -ln[Z] -0.105 ln[t90](2.53) =-

i.e. a plot of ln[tgo] versus 10001Twill be linear and the slope will beEa/R, where Ea will be in kCal (rather than in cal) per degree per mole.

Extrapolated Curve: y = 4.6Q~-Q.OQl36t

4.61

E x p e ~ ~ e n tPoints ai Obtained Later

4.60 4.59 4.58 4.57

4.56 4.55 0

Time (months)

2 Extrapolated decomposition curve of a solution obtained (e.g., in January 1993 after three months accelerated data) compared with the actual data accumulated in time and plotted 33 months later (i.e., in October 1995). able

Data from Table 6 Plotted by the

t90

Method $90

Temperature 55°C 45°C 37°C a

1000/ T 3.049 3.145 3.224

kl (mo") 0.031 0.01 18 0.0052

(months)a

ldt901

3.40 8.93 20.26

1.233 2.189 3.009

Calculated fromkl values from the equation kt90 = -0.1054.

Table 8 shows the data from Tables 6 and 7 plotted by the $90 method. The data are shown in Fig. 13. The least squares fit equation in Fig. 13 is ln[tgo] = -29.695 (10.142/ Q, and it is noted that the slope (activation energy) is the same as in Fig. 11 (where the slope is 10.153). The small difference lies in rounding off errors in calculating t90 from the kl values. The t90 value at room temperature is determined to be 4.072 [Eq. (2.53)]. If plotted on semilogarithmic paper, the valuewillemerge directly, and this type of plotting is more easily understood by those not familiar with kinetics than plots of the type in Fig. 11.

One advantage of testing at higher temperatures is that it is possible to construct decomposition profiles at nonconstant temperature. It should be pointed out that extended room temperature is defined in the USP, 1990, as between 15 and 30°C but isnow15-25°C. Stability studies are usually carried out isothermally, e.g., as mentioned, at 25°C. The rationale for testing temperatures willbe discussed at a later point in the book; suffice it to say at this point that a product in the marketplace will never experience an isothermal shelf history.

3.0

3.1

3.2

1ooo/r

Arrhenius plotting using

3.3

3.4

tgo.

At best there willbe daily fluctuations. This concept was investigated by Carstensen and Rhodes (1986), in the following fashion. Suppose a product is stored at 25°C with daily fluctuations of f5"C. This means that the dependence of T on time, t (in days, if the cycle is one day), is given by

T = TI (2.54) + T2 sin(2n;t)

This is shown in Fig. 12. Assume, as well, that the reaction is zero order, i.e.

C = CO-kt

(2.55)

Introducing Eqs. (2.50) and (2.54) into this and applying the situation to a differential time element dt, gives (2.56) so that to obtain the concentration after a time period t of cyclic storage, where 0 t 1 day, is givenby (2.57)

These types of integrals can be solved by computer programs. The loss in e.g. one year will be 360 times that after one day, and so on. The authors calculated the loss after 3 years of storage, using di vation energies, and using k25 (isothermal) =0.01%, and using a daily cycle with a fluctuation of f5"C. They arrived at the results in Table 9, It is seen (Table 9, Fig. 14) that as long as the activation energy is less than 22 kCal per mole,the percent increase in the amount lost after a given storage period is lessthan 10%. For example, if a dosage form lost 5% after 3 years storage at static room temperature, it would lose 5.5%after 3 years of cyclicroom temperature. This point will be of importance in the following section.

Cyclic Versus Constant 25°C Data. kzs =0.01% Per Day E, kCal 3 moleper

Loss after years

1.910 3.9 15 20 12.1 25 17.7 30

11.16 11.38 11.77 12.27 12.89

Percent increase loss in

7.5

Hours

3 6 48 42

Daily temperature fluctuations according to Eq. (2.54).

It is noted that any type of cyclecan be used, e.g.seasonal cycles could be used as well. The problem of cyclic testing (for chemical stability) is raised from time to ut to decide on a cycle is difficult, and it is much more rational to use the data from accelerated studies to produce the desired profile. is

It is possible,rather than studying a reaction at a fixed temperature (isothermal~y),to vary the temperature in a given fashion, and fit the data to Eq. (2.50). For a zero-order reaction we can write CCO = kt = [ z e . p ( - ~ ) ] .

(2.58)

e may allow T to vary in a given manner, e.g., in the simplest case as

1 T

-=a-bbt

(2.59)

where a and b are the constants that we input into a programmable temperature

Alkaline Decomposition of Riboflavin by Isothermal and by Several Nonisothermal Programs Rate constant at 25°C

Temperature program Isothermal

0.016

Linear up Linear up Log UP Log upa Log down

0.014 0.016 0.018 0.015 0.015

Activation energy mole) (kcal/Reference Cole 19.2

and Leadbeater (1 966) 20.1 Guttman 962) (1 20.3 Rosenberg al.et (1984) 17.9 Madsen et al. (1974) 20.9 Rosenbergal.et (1984) 18.9 Rosenbergal.et (1984)

Triplicate experiments. Table constructed from data published by Rosenberg et al. (1984).

Source:

y inserting Eq. (2.59) into Eq. (2.58), the concentration profile in time becomes

(2.60)

C or, logarithmically,

[~l

Ea ln -= l n [ a --(a

-bt)

~]+

= [ I n ( Z ~-

(2.61)

(2)t

This gives rise to a linear plot when ln[Co/q is plotted versus t, and (since a and b are the constants for the program we have chosen for our temperature bath), the slope (divided by blR) will give us the value of Ea;and now l n [ a can be obtained from the intercept. The same procedure can be used for first-order reactions, although they are somewhat more complicated. In such a case a computer program is best,and curves can be generated to match the curve obtained experimentally, Table 10 shows someof the investigations that have beencarried out in the last 20 years, using this procedure.

Since stability studies carried out in an industrial setting are isothermal, there has been a fair amount of discussion over the last 2 decades as to what the actual temperature of the study ought to be. Up to 1993, the FDA required stability studies to be carried out at 30°C for the approval of expiration periods. In contrast, the European comunity would consider the United States as an area where 25°C would be the appropriate temperature to require for isothermal testing. The resolution of the problem in Europe came from by Futscher and Schumacher (1972), who established the climate zones,and Wolfgang Grimm (1985,

25 20 a,

E I“

15 10 5 0 0

l 0 1 1 1 2

Time

Hypothetical temperature profile a drug might experience in noncontrolled conditions in a temperate climate.

1984),whoemployed the concept of virtual temperature (Haynes, 1971) to systematize them. The zones are as in the table. Temperate Mediterranean Tropical moist Desert

21°C (45% 25°C (40% 30°C (70% 30°C (35%

RH) RH) RH) RH)

The temperatures listed are not the average yearlytemperatures in theseareas but rather the kineticmean temperatures (KMTs). These are obtained by the following considerations. What is important is not what the temperature is, but how much drug is lost. The problem inthe United States is that certain areas (for instance Arizona) are in the desert zone and certain areas (e.g., Miami, the Keys) are in the tropical zone. If a product is shipped to any of these areas it may experiencetemperatures or relative humidities other than those of Zone 11. Consider a drug product that experiencesthe temperature profile in Fig. 15. If a sample experiencesthis temperature profile, then the rate constant will first increase from a low in January to above the “average” in the summer, then fall back to the “average” rate constant right after the summer, and decreasebelow the “average” as the fall months set in. The actual concentration profile is the curve shown in Fig. 14. If the average temperature had been used, isothermally, then the line in the figure would have resulted. It is seen that if the average temperature from Fig. 15 had been used, then the fraction of drug retained after 12 months’ storage would have been higher than in the case of the nonisothermal storage. There is a temperature, which, had the drug been stored at that temperature isothermaly, would have caused a drop exactly equal to that brought about by the nonisothermal storage. This is shown in Fig. 17. The question is, What is that temperature?

l .02 1.oo

ClCo in Variable Temp Storage

0.98 0.96 0.94 0.92 0.90 6

0.88 40

Sept July March May Jan

Nov

6 Nonisothermal temperatureloss using thetemperature profile in Fig. 15,compared with loss encountered by isothermal storage at the average temperature from Fig. 15. l .02 ClCo in Variable Temp Storage

1.oo

0.98 0.96 0.94 0.92 0.90

C/Co at Kinetic Mean Temperature

0.88 40

Sept July March May Jan

Ti me

Nov

17 Definition of kineticmean temperature as the temperature at which the loss, isothermally, equals that of the variable temperature storage.

So for this approach it is necessaryto know (a)the absolute temperature (To profile as a function of time, t:

and (b) the activation energy. If we do, we may calculate the amount retained after, say, 12 months as: (2.63)

The kinetic mean temperature (Tkmt) is then defined as that temperature at which isothermal storage (at Tkmt) would have causedthe same drop in potency, i.e., (2.64) It is noted that 2 cancels out, so that knowledge of the pre-exponential factor is ence the kinetic mean temperature for the reaction in ~ ~ e s ~isi o n given by (2.65) The approach of G r i m for a harmonized, global approach was to assume that (as has been reported, e.g., by Kennon, 1964) that on the average E a = 10kCal/mol R

(2.66)

Tkmt with this restriction is calledthe kinetic mean temperature (KMT), and the only requirement for its calculation is knowledge of the temperature profile of storage. Grimm (1985, 1986) calculated the KMT for a series of cities and countries and arrived at the climate zones listed above. As a result of an FDA advisory board meeting (6/25/ 1993, Silver Springs, and the following ICH Guidelines, the agency has adopted this polic for nited States as well. This is shored up by the so-called Prescription Drug keting Act( P D M ~ )which , requires that all warehouses, distribution centers, transThat is, the portation chains, and pharmacies retain KMTs oflessthan.25°C. average measured temperature should be below 25°C and only “spurious excursions” to 30°C are allowed. It is noted that the average measured temperature is always (except in truly isothermal situations) lower than the KMT. In actual isother~aZstorage, the KMT is equal to the temperature of the storage.

Recalling the thermodynamic relationship AG

-TAS

(2.67)

the Arrhenius equation can be written (2.68) Eyring has shown that

z=-KT h

(2.69)

~rstensen

where

is Boltzmann’s constant,

h: = 1.38 *

ergldeg

(2.70)

and h is Planck’s constant,

h = 6.63

erg S

(2.7l)

hence (2.72) or l n [ ~ ]= 23.76

(2.73)

Introducing this into Eq. (2.68) then gives (2.74) or (2.75) where the superscript refers to the transition state (Laidler, 1965). Kearny et al. (1993) employed these principles for interconversion kinetics and equilibrium of the lactone and hydroxyacid forms or the HMG-CoA reductase inhibitor, R*)]-2-(4-fluorophenyl)-b,d-dihydroxy-5-( l-meth~leth~l)-3-phen(21-981 = [R-(R*, yl-4-[(phenylamino)carbonyl]-1H-pyrrole-1-heptanoic acid. The equation they arrived at was

AS*

AH*

1 n [ ~ ]=31.95+--1.g87 1.9871”

(2.76)

If a drug substance A reacts with a second substance B according to a common scheme: A+B-+C

(2.77)

then the rate equation is

--k2[A][B]

”

(2.78)

If the initial concentration of A present is a,the initial concentration of I3 present is b, and the amount of C formed at time t is x, then at time t

[A] = a -x

(2.79)

[B] = b -x

(2.80)

and

Inserting this in (2.61) gives dx -= -k2(a -x)(b -x)

(2.81)

which integrates to l a(b n [ x)~ =] (a -b)k2t

(2.82)

or l n [ ~ =] (a -b)k2t

+ In [~l -

(2.83)

It is notedthat the term on the left-hand side of the equation is dimensionless, so that the right-hand side is so as well. Hence the concentrations, a, b, and x may be in anyconcentration unit, and the unit for k will be suchthat the right-hand side is dimensionless, i.e., [time concentration]". To establish the general shape of the curves of a, b, and x, and second-order rate curve has been generated using A(0)= a = 100 and B(0) =b =70 and k2 =0.0001 This curve is shown in Fig. 18.

Y-

Time Generated second-order reactioncurve. Concentration of reactant (decomposition product) is plotted versus time. a = 100, b =70, and k2 =0.0001.

Second Order Data Where a =A(0) = 100, Assuming Values of b =B(0) = 60 and 70 x

24 31 69 63 53 47

Time 0 2 56 4 6 39 8 43 10 12 23

0 14 37 42

[B]=60--x

[B]=70-x [A]=G-x

60 46 36 29 33 18 13

100 86 76

A set ofsecond-orderdate areshown in Table 11, and it is assumedthat A, with an initial concentration of 100, reacts with an unknown substance, B, i.e. that b is unknown. a and b couldbe, for instance in mg/mL and k could be in units of mL(month - mg). The number of reacted concentration units (x) isshown graphically in Fig. 18 as a function of time. (i.e., the ordinate could be in mg/mL, and the abscissa in months). If the task were to study the reaction between A and B, then the experimentaldesignwouldsimply be to carry out the reaction with varying amounts of A and B and plot these according to Eq. (2.82) or (2.83). The problem in stability work, however, is quite different. In general the investigator follows the concentration of active drug, say A, and does not know that an interaction is taking place. (It is not possible, a priori, to imagine all the possible interactions that may occur.) An experience of the author’s is a parenteral, aqueous diluent that contained maleic acidand propylene glycol.In general, one expects that anester will hydrolyze, but rarely that itwill becreated by the opposite reaction in an aqueous system. In this case, however, there was a formation of both the mono and diester of propylene glycol and maleic acid. This was eventually deduced, but it was not thought of originally, With good assay methodology, this may be discovered early, but not necessarily, because,if it is not suspected, the assay may not account for the possibility of the interaction, and the reaction product peak can hide under a reactant peak, if HPLC is used. An example is shown in Table 12, where concentration is monitored versus time for progabide in pH 1.75 buffer, This decomposition has been reported by Farraj et al. (1988). The data in Table 12 are the same as those used to create Figs. 8 and 9. Suppose, first of all, that itwere not suspected that the decomposition of the drug A was due to interaction. Then the logical procedure wouldbe to plot the data first order (semilogarit~ically),and, in this vein, the data from Fig. 18 are plotted in Fig. 19. Note that the terms (a-x) and (b -x) are the concentrations of unreacted (undecomposed, intact) A and B, respectively, but that since b is not known, then neither is x nor (a-x) nor (b -x).

n.)

le 2.12 Decomposition of Progabide in pH 1.75 Buffer

as a Function of

Time Concentration Time

0 7 13 19 25 31

C 0.00896 0,00760 0.00700 0.00660 0.00612 0.00600

ln[C-0.00551 [C-0.0055] 0,00343 0.00205 0.00245 0.00105 0.00062 0.00045

-4.71 5 -4.880 -4.962 -5.021 -5.096 -5,116

Source: Data taken off graph published by Farraj (1988). et al. Table numbers are subject to themagnitudeoferrorassociatedwithreadingoffgraphs.Thedataareusedina pedagogical sense.

The manner in which this is firstapproached is by iteration. Several valuesfor the niveau value, Cinf, are selected and plotted by the equation

The plot using Cinf =0.0055 is shown in Fig. 8. This type of process is usuallycarried out by computer by inserting different values ofCinf into Eq. (2.84) and selecting the “best” value by the criterion that the produced line gavethe smallest valueof the sum of squares (S&, to be discussed in Chapter 11, “Statistical Aspects”). It is seen in Fig. 19 that there is a “trend in the data,” i.e. that the plain first-order plot is not linear. It is not just that the correlation coefficient is rather small* but it is the fact that the points collectbelow the linein the middle region. If indeed this were an interaction, then one might guess that there would be a concentration where there would be a leveling off (from Fig. 5 an infinity value Cinf =0.0055 would seem reasonable). If this were a second-order reaction, then this concentration, with concentrations expressed in molarities, would be the value of a -b. It isknown,now, that the value of b in the above exampleis CO-Cinf =0.00896 -0.00555 =0.00341 molar. It is not known what B is, but the concentration of it is known. Thiscan now be usedto plot the data by Eq. (2.82) or (2.83) as has been done in Fig. 20. The intercept is not quite in line with what it should be, since ln[O.00896/0.00346]= 1.24, but it is sufficiently close to assume that the treatment is in the right direction. A s mentioned, computer programs are the easiest way of handling second-order reactions, and a program for data treatment of second-order kinetic data is shown here.

* The correlation coefficient is close to 1.00 for very linear data. coefficient for six data points.

0.97 is not too good a correlation

arstensen

-5.0 -5.1

In[C]=? -5.2 0

Time (min) Data fromTable11plottedsemilogarithmically.The

least squares fit is

y = -4.764 -0.0127~(R=0.97).

0 0

Time (min) by Eq. (2.82) using a =0.00896 0.051~.

The data in Table treated 4

b =0.00896 -0.0055 =0.00346. The least squares fit is y =0.92

and

= ";W1 l00 INPUT "INITIAL DRUG CONCENT~ATION 105 INPUT "INITIAL REACTANT CQNCENTRATION = ";W2 l10 INPUT " N U ~ B E ROF POINTS = ";N1 115 PRINT "PRINT DATA IN 400 BLOCK AS: DATA,CONC ACTAN ANT" 120 PRINT 125 PRINT t'TIM~tt;SPC(6);''CONC.~ECO~P.'t;SPC(6); ltLN~B(A-~)/(A(B-~))~" . , . . . - ., .- ..'l 130 PRINT It- - - - .= -. 200 READ B,A 205 X = A 210 W 3 = W - B 215 W4 =W2-B l

olution Kinetics

220 W5 =w 2 / w 1 225 W6 =W5*W3/W4 230 Y =LOG(W6) 235 X1 =X1 + X 240 X2 =X2 (X"2) 245 Y l = Y 1 + U 250 Y2=Y2 (Y2"2) 255 Z1= Z1+ (X*Y) 260 N2=N2+1 265 PRINT X;SPC(S);Y 270 IF N2 = N l GOTO 700 300 GOTO 200 400 DATA 1.32,6 405 DATA 2.93,12 410 DATA 4.96,18 415 DATA 7.59,24 420 DATA 11.2,30 425 DATA 16.5,36 430 DATA 25.8,42 435 DATA 5113.48 700 Z2 =X2 -((XlA2)/N2) 710 Z3=Y2-(YlA2/N2) 720 Z4=Z1 -(Xl"YlIN2) 730 z 5 =z41z2 740 Z6 =(Y1= (Z5*Xl))/N2 750 Z7 =(Z4"2)/(23*22) 760 78 =Z7yO.5) 770 PRINT 800 PRINT "SLOPE = ";Z5 810 PRINT 820 PRINT "INTERCEPT = ";ZL6 830 PRINT 840 PRINT "CORRELATION CO~FFICIENT= ";Z8 850 PRINT 860 PRINT "S"2= ";Z9 870 END

+ +

Second-order reactions, per se,are less frequent in the literature, because often the reactant is unknownin stability situations. Ohta et al. (1998) found that oxidized glutathione (GSSG) interacted with trans- and c ~ s - d i a ~ ~ i n e d i c ~ o r o p l a t i n u ~ (DDP) by way of the S-S bond. The reaction was pseudo-first order and the rate constants linearly related to cis-DDP concentrations. If one considers the reaction GSSG

+ DDP + products

then, if the GSSG concentration is high,the reaction is secondorder but will appear first order and, indeed, the pseudo-first order rate constants will be linear in DDP concentration.

l~o~centr~tio~,

In the case where a =6, the rate equation becomes (2.85) This may be written d(a -x) --k2t (a -x12

(2.86)

which integrates to

[ ~ [~] ] =

+k2t

(2.87)

It should be noted in the foregoing whether one of the two components A and B is >[A], then the rate equation becomes much less than the other. If [B] >

dA = -k2[A][B] dt

-kps[AI

(2.88)

where kps = k2[331

(2.89)

kps isa pseudo-first-order rate constant, so that in this case the decomposition will appear first order in the predominant species. Particularly where kps= k2 55.5 since 55.5 is the molarity of water in water, and since it is assumed >[A], so that the final concentration of B is that [B]>

If b >a, then the integration of Eq. (2.81) gives (2.91) If, however, b >>a, then (2.78) becomes

--~k~[B]}[A~ M -K,[A]

(2.92)

where k2PI = %

(2.93)

is almost constant and is denoted the pseudo-first-orderrate constant. It is noted that this will have the unit of time as it should. Most reactions actually are of this type. A common case in point is that of a hydrolysis. The molarity of pure water is

10001 18= 55.5. An example is the case of hydrolysis aspirin of (MW 180) (Edwards, 1950):

If the solution of aspirin were, e.g., 0.1 molar at the onset, then ifthe reaction went to completion, the final water concentration would be 55.4 molar, i.e., would hardly change at all. Hence most hydrolyses are pseudo first order. The literature isrepletewithexamples of pseudo-first-orderhydrolysis. Lazarevski et al. (1978) reported the kinetics of the acid catalyzed hydrolysis of erythromycin oxime and erythromycylamine to be pseudofirst order. Sternson and Shaffer (1978)reported the hydrolysis of digoxin in acid solutions to be pseudo first order, although the reaction scheme was complex. Hydrolysis is usuallythought of as being undesirable. However,in the case of prodrugs, it is a built-in feature: prodrugs are utilized if the parent drug is poorly soluble (Le., making a soluble derivative)or poorly bioavailable. In the proper biological environment, it should decompose into the active species (e.g., hydrolyze in the bloodstream after absorption). The point is that A is a derivative of an active species which isnot absorbed well, and that A itself is well adsorbed. It must then (unless it too is active) decomposeinto the active speciesin the bloodstream so that the action is indistinguishablefrom that of the parent compound. Vaira et al. (1984) havedescribed the kinetics, pH profile, and Arrhenius plots of derivatives of 3-[hydroxymethyl]-phenytoinesters. It was found that various amino groups containing acyl esters of 3-(hydroxymethy1)phenytoin were potential orally useful products and that the disodium phosphate ester of 3-[hydroxymethyl]-phenytoin appeared to be a good parenteral form of the drug.

Frequently it is assumedthat since a reaction is a hydrolysis, then, the scheme should be pseudo first order. Although this may always be true in the initial stages of a decomposition, caution should beobserved in extending the conclusions.An example isthe work by Teraoka et al. (1993) dealing with hydrolysisat the S-bond of ranitidine in acetate buffer and in water. Their findings are shown in Fig. 19. When such occurrencestake place, more sophisticated mechanisms mustbe thought of (e.g., Franchini and Carstensen, 1994). Teraoka et al. (1993) in the abovecase established that a set of parallel reactions (hydrolysis, two rearrangements) occurred, i.e., denoting ranitidine by I and other products by higher roman numerals, the scheme proposed was I "+ I11 I -+ I1 I -+ IV H20 (ring closure) I1 I + condensation

Of these the first three will befirst order (parallel), and hence, often in stability work, a first order assumption is not unreasonable (Teraoka et al., 1993) (see Fig. 21).

cn c: .._. e:

~rste~sen 100 80

60 +

100

200

Time, hours ig. 2 Ranitidine stability in acetate buffers (0.1 M) of ionic strength 0.1. (Graph constructed from data published by Teraoka et al., 1993.)

Angerg, M., Nystrom, C., Castensson, S. (1988). Acta Pharm. Suec. 25:307. Angerg, M., Nystrom, C., Castensson, S. (1990). Int. J. Pharm. 61:66. Angerg, M,, Nystrom, C., Castensson, S. (1993). Int. J. Pharm. 90:19. Archontaki, H. A., Gikas, E. E., Panderi, I. E, Ovszikoglou, P. M. (1998). Int. J. Pharm. 167:69. Aso, Y, Sufang, T., Yoshka, S., Kojima, S. (1997). Drug Stability 1237. Beal, H. D. Pranderd, R. J., Todaro, L. J., Sloan, K. B. (1993). Pharm. Res. 10:905. Beal, H. D., Prankerd, R. J., Sloan, K. B. (1997). Drug Dev. Ind. Pharm. 23517. Burke, M., Redden, P. R., Douglas, J.-A,, Dick, A., Horrobin, D. F. (1997). Int. J. Pharm. 15791. Buur, A., Bundgaard, H, (1984). Int. J. Pharm. 21:349. ' Carstensen, J. T. (1997).Modeling and Data Treatment in the Pharmaceutical Sciences, Lancaster, PA: Technomic Publishing Corp. Carstensen, J. T., Attarchi, F. (1988a). J. Pharm. Sci. 77:314. Carstensen, J. T., Attarchi, F. (1988b). J. Pharm. Sci. 77:318. Carstensen, J. T., Franchini M. K. (1994). Pharm. Res. 11:s-26. Carstensen, J. T., Rhodes, C. T. (1986). Drug Dev. Ind. Pharm. 12:1219. Carstensen, J. T., Su, K. S. E., Maddrell, P., Johnson, J. B., new mar^, H. N. (197 l), Bull. Parenteral. Drug Assoc. 25: 193. Carstensen, J. T.,Attarchi, F., Hou, X.-P. (1985). J. Pharm. Sci. 74:741. Cole, B, R., Leadbeater, L. (1966). J. Pharm. Pharmacol. 18:lOl. Edwards, L. J. (1950). Trans. Faraday Soc. 46. Fabre, H., Eddine, N. H., Berge, G. (1984). J. Pharm. Sci. 73:611. Farraj, N, F., Davis, S. S., Parr, G. D., Stevens, H, N. E. (1988). Pharm. Research 5226. Franchini, M. K. (1995). Ph.D. thesis, University of Wisconsin, p. 158. Franchini, M. Carstensen, J. T. (1994). Int. J. Pharm. 111:l 53. Franchini, M. K., Carstensen, J. T. (1999). Pharm. Dev. Tech. 4:257. Futscher, N., Schumacher, P. (1972). Pharm. Ind. 34:47. Ghebre-Sellassie, I., Hem, S. L., Knevel, A. M. (1984). J. Pharm. Sci. 73: 125. Grimm, W. (1985). Drugs Made in Germany, 28:196. Grimm, W. (1986) Drugs Made in Germany, 29:39. Guttman, D. E. (1962). J. Pharm. Sci.51:811. Hammad, M. A., Muller, B. W. (1998). Int. J. Pharm. 169:55,

Hansen, L. D., Lewis, E. A., Eatough, D. J., Bergstrom, R. G., DeGraft, K., Johnson, D. (1989). Pharm. Res. 6:20. Hashimoto, N., Tasaki, T., Tanaka, H., (1984). J. Pharm. Sci. 73:369. Haynes, J. D. (1971). J. Pharm. Sci. 60:927. Higuchi, T., Rheinstein, J. A. (1959). J. Am. Pharm. Assoc., Sci. Ed. 48:155. Jordan, C. G. M. (1998). J. Pharm. Sci. 87:880. Kabadi, M. B., Valia, K. H., Chien, Y. W. (1984). J. Pharm. Sci. 73:1461. Kearney, A. S., Crawford, L. F., Mehta, S. C., Radebaugh, G. W. (1993). Pharm. Res. 10:1461. Kennon, L. (1964). J. Pharm. Sci.53:1067. Laidler, K. J. (1965). Chemical Kinetics, New York: McGraw-Hill, p. 68. Lazarevski, T., Radobolja, G., Djokic, S. (1978). J. Pharm. Sci. 67: 1031. Lerner, D. A., Bonneford, G., Fabre, H., Mandrou, B., deBuochberg,M. S. (1988). J. Pharm. Sci. 77:699. Li, S., Wang, W., Chu, J., Zamansky, I., Tyle, P., Rowlings, C. (1998).Int. J. €%arm. 167:49. Madsen, B. W., Anderson, R. A., Herbeison Evans, D., Sneddon, W. (1974). J. Pharm. Sci. 63:777. Malkki-Lame, L., Valkeila, E. (1988). Int. J. Pharm. 161:29. Misra, P. K., Haq, W., Katti, S. B., Mathur, K. B., Raghubir, R., Patnaik, G. K., Dhawan, B. N. (1993). Pharm. Res. 10:660. Nguyen, T. H., Burnier, J., Meng, W. (1993). Pharm. Res. 10:1563. Notari, R. E., DeYoung, J. L. (1975). J. Pharm. Sci. 64:1148. Ohta, N., Inagaki, K., Muta, H., Yotsuyanagi, T., Matsuo, T. (1998). Int. J. Pharm. 161:15. Oliyai, R., Lindenbaum, S. (1991). Int. J. Pharm. 73:33. Owen, W. R. Stewart, P. J. (1979). J. Pharm. Sci. 68:992. Prankerd, R., Walters, J., Parnes, J. (1992). Int. J. Pharm. 7859. Rosenberg, L. S., Pelland, D. W., Black, G. D., Aunet, D. L., Hostetler, C. K. (1984). J. Pharm. Sci. 73:1279. Shalaev, E. Y., Shalaeva, M., Byrn, S. R., Zografi, G. (1997). Int. J. Pharm, 152:75. Sternson, L. A., Shaffer, R. D. (1978), J. Pharm. Sci. 67:327. Teraoka, R. Otsuka, M,, Matsuda, Y. (1993). J. Pharm. Sci. 82:601. Varia, S. A., Schuller, S. Stella, V. J. (1984). J. Pharm. Sci. 73:1074. Vilanova, B., Mutloz, F., Donoso, J., Frau, J., Blamco, F. G. (1994). J. Pharm. Sci. 83:322. Visconti, M., Citerio, L., Borsa, M., Pifferi, G. (1984). J. Pharm. Sci. 73:1812. Yip, Y. H, W., Po, A. L. W., Irwin, W, J. (1983). J. Pharm. Sci. 72:776,

This Page Intentionally Left Blank

adi is on, Wisconsin

1. Ionic Strength 1.1.TheDebye-Huckel and subsequent limiting laws 1.2.p&,p&, andpK 1.3. pH and the Henderson-Hasselbach equation 1.4. Ionic strength of mistures of weak and strong electrolytes

59 59 60 62 65

2. General and Specific Base Catalysis. Buffer Effect 2.1. Is the optimum pH the same as the pK? 2.2, Temperatureeffects 2.3.Acid or base catalysis

67 71 72 73

Profiles for the Situation Where HA

+ H+ and

+ H+ -+ products 3.1. Specific acid/ base catalysis of one species (HA + H+ and HA. + OH" -+ products)

A.-

75 77

4. Hydrolyses of Acidic Drugs at a Given pH

5. A Simplified Approach to pH Profiles

6. Effect of Buffer

7. General and Specific Acid and Base Catalysis

8. The General Procedure

Profiles in Stability Work 9.1. SubtypeA 9.2.SubtypeABC 9.3.SubtypeABCI) 9.4.SubtypeABCDE

86 87 88 88 91 57

9.5. 9.6. 9.7. 9.8. 9.9. 9.10. 9.11. 9.12. 9.13. 9.14. 9.15. 9.16. 9.17. 9.18. 9.19. 9.20. 9.21.

Subtype ABCDEF Subtype ABCDEFG Subtype ABCDENOP Subtype ABCDEFGPQT Subtype AH Subtype AHJ Subtype AHJK Subtype AHJKL Subtype AHJD Subtype DENOPQ Subtype AHJDEF Subtype BCD Subtype HJDE Subtype HJDEF Subtype H J K L ~ Subtype JDEN Subtype JDENOP

92 93 94 95 95 96 97 97 98 98 99 99 100 102 102 102 102

10.Effect

of Concentration

103

11.Effect

of Solvent

103

12. Kinetic Salt Effect

104

References

107

One of the tasks of stability scientists, particularly in the preformulation stage, is to establish the effect of pH on the stability of the drug. To this end it is worthwhile to describe a couple of concepts in wideuse,concepts that are often misused. These concepts are closely tied to the effect of pH on reaction kinetics. It is, therefore, important to ascertain a “constant” pH during the reaction, and there are two means of doing this: (a) using a pH stat and (b) employing buffers. In the former case, the “salt” concentration changes only verylittle, and one assumes (maybe not justi~ably)that the determination is carried out “in water.” In the latter case, there is the complication that the buf€er species (and there are always at least two for the buffer to perform its function) catalyzes one or more must be taken to this effect. of the reactions, so that steps The second point is that the mere presenceof an electrolyte (NaCl for instance) can affect the rate constant and the pK values in question and that, therefore, a complete kinetic study must also address this point. The important variables in a kinetic study of a drug substance in solution are therefore Buffer effects Ionic strength effects PH pK of drug substance

The ionic strength, p , is defined as

where mi is the molality of the ith ion and Zi is the charge of the ith ion. A word of caution is in order. A “salt” is considered, in this text, to be the solid state of a strong electrolyte that, in solution, is completely dissociated, e.g., calcium chloride:

For such a 2 :1 electrolyte, if the molality of the “salt” were m molal, the molality of the chloride ion would be 2m molal, so that these figuresinserted into Eqs. (3.1) and (3.2) give p

= 0 . 5 [ ( ~X 22) (2m X 12)] 3 M

(3.3)

and this is true for a 2 :1 and 1 :2 strong electrolyte in general.For a l :1 electrolyte, p=m.

It should be added at this point that manyinvestigationsdealwithconcentrations, where, in fact activities are the important variable. On the surface, it should therefore be a simple matter to calculate the ionic strength of a preparation. When, however,the case is one of weak electrolytes (acetic acid, acetic buffers), the problem becomes more complicated. The degree of dissociation of an acid can for instance be calculated via the acid equilibrium constant, aswillbe covered shortly. This leads to a degree of dissociation, a, so that the ionic part of the solution contributed by the weak electrolyte is, e.g., in the case of a simple acid (HA),

p=ma

(3-4)

The problem with determining the degree of dissociation will be dealt with below. The activity, a A , of a compound is related to its concentration [A] by the activity coefficient f~ (Moore, 1963):

It will be shown later that the ionic strength can have a definite effecton stability of drugs. van Maanen et al. (1999) havedemonstrated a kinetic salt effect in the aquous solution decomposition of thiotepa. 1.1.

The D e b y e - ~ ~ c k ean l Subsequent

limit in^ Laws

The activity coefficient, f, of a compound is the ratio between activity and concentration, e.g., for a compound HA,

where [HA] is the concentration of HA in molality. Often, particularly in more dilute solution, there is a tendency to replace molalitywith molarity, but it is a good habit

to dowork in units of molality, in particular because molalitiesare not temperature at 50"C, and this statement dependent; a one molal solution at 25°C is still one molal cannot be made for a one molal solution. The Debye-Huckel limiting law states that -logf, = 2A- JP (&133/2

(3.7)

where z is charge, E is dielectricconstant, T is absolute temperature and log isa base 10-logarithm (Brmnsted, 1943). For aqueous s o € ~ t i o(E ~= s 80) at room temperature (25"C), the factor

and this factor is often used indiscri~inately at other temperatures, and sometimes in nonaqueous media, so

This is true at veryhigh dilution. In slightly more concentrated solutions, the equation takes the form (3.10) where the factor Pis related to the ionic radius of the electrolyte and is often set to be one, i.e. (3.1 1)

These three concepts are frequently bandied about and described indiscriminatelyin literature, and definitions of these (Brmnsted, 1943) are quite old, but the laws of thermodynamics, we hope, are notfunctions of time. (The notation H+ rather than H30+ will be used in the following.) The ionization constant of an acid, based on the equilibrium

is (3.12) The notation of Brmnsted is used here. The where brackets denote co~ce~trations. It is most often simply term "operational bpK" is used and it will be written K(&).. denoted K. It will be seen in the following, that simply using the term K can be

misleading, sincethen it would be difficultto distinguish betweenthe same concept for a base. The pK(,) is defined as the negative of the base-ten logarithm of K, i.e.

y ~ a ~ iconstant ~ is defined as The t ~ e r ~ o ~equilibrium

(3.14) i.e. there is no parenthesis in the subscript. (Actually the Bronsted notation was squared brackets, but the term Ka, thusly written, is so commonplace that it is used here without brackets). It follows that

Since activitiesare connected with concentrations by way of activity coefficients,f, e.g., aH+ =fH+[H+]

it follows that

(3.17)

If the charge of HA is ZHA (i.e., would be 0 for acetic acid, but would be 1 for an amine hydrochloride), then that of A-, is

and since that of the hydrogen ion is

= -PK(~)-0.5[12

+ 1 it follows in general that

+ ( 2 ~ "-)(ZA~ + 1)2],/p (3.19)

This is identical to the formula quoted by Brarnsted (1943) where log K(a) rather than PK(a) is used, hence there is a plus, not a minus, in front of ZA-.It is noted that it is the charge of the conjugate base, whichis the Z value, that must be used.

It is noted that the PT((,) is the value that must beused in a particular systemwith a particular ionic strength. It ishence a function of the system, and therefore, if tabulated, wouldhave to be referred to in relation tothat specificsystem. In contrast to this the pKa (and pKb) are universal quantities that maybe tabulated. To obtainthe operational PIS(,), the value for the pK, and the value of ZA-,/p are inserted in Eq. (3.20). To obtain J p , when a weak acid is in question, the Henderson-Hasselbach equation, cited in the following section, is employed. This allows for calculation of the degree of dissociation, a. As will be seen below, this may require a further iteration, since the new valueofwillgive a different degree of dissociation, and hence a different ionic strength, than first employed in the equation. pK values can be obtained spectrophotometrically. For instance, Sing et al. UV spectrophotometric methods to determine the p& of (1999)haveused nimesulide. e~ders~n- asse el

The hydrogen concentration of a solution is usually described by its base-ten logarithm and is then referred to as the pH (Smensen, 1909a,b, Moore, 1963). That is, (3.21) where [H"] is the hydrogen ion conce~tration(not activity). Bates (1964), however, distinguished between (3.22) and (3.23) In this writing, the definition in (3.21) will beused, so that care should be taken in the comparison of expressions derived by the second definition of pH. With the definition in Eq. (3.21) the equilibrium equation in Eq. (3.1 1) can be written (3.24) This assumes that the activity coefficients are unity. It is usually the total concentration, CO,of"acid" that is known, e.g., one talks about a solution that is one molal in acetic acid. Hence (3.25) and this inserted into Eq. (3.24) now gives (3.26)

Hence knowing the "concentration" COand the pK(a) it is possible to calculate the amount of [A-] and [HA] at any pH. Water is a special case: H20 -+H'

+ OH-

so the concentration of[H'] and L0.H-J are each product of water is given by Kw= [H'][OH-]

(3.27) at pH 7, and the ionization

(3.28)

pKb is the quantity for a base corresponding to pK,. For a base the reaction is (Martin et al., 1993): B"

+ H20 -+ HI3 + OH-

(3.29)

and it follows that (3.30) Multiplying and dividing Eq. (3.30) by Eq. (3.28) and employing Kw = [H+][OH"] now gives (3.31) The relation of K(b) to Kb is completely analogous to the relation of K(a) to Ka discussed above. Knowledge of PKa and ksb values can be used roughly to estimate the pH of solutions of the drug substance. An example is shown here. Example 3. l.

Vecuronium bromide (Merck Index) has a PKb of 9. The molecular weight is 637, and if a solution is 0.9%, what would be the pH of such a solution? Answer.

The solution would be 9/ 637 =0.014. molar, In this case the equilibrium is RH'

+ H20 -+RH;'(3.32)+ OH"

To calculate the pH of a 0.9% solution of the quaternary salt in aqueous solution, it is first necessary to establish what the ~ K ( Bis.) It is hence necessary to calculate the ionic strength. To do this one would have to know how much of the vecuronium was present as monopositive and dipositive ion. To do this one would haveto know the ~ K ( B )i.e., , onewouldhave to know the quantity one seeks. The problem can be solved by continuous iteration.

As a first guess, one would use the pKb value and assume it to be the pK(b) value. The pH is then given by artin in et al., 1993):

(3.33) which gives POW = 5.43

(3.34)

or (3.35) 5.43 = 8.57 pH = 14 artin reference isa basic text and does not go into ionic strength effects, but it is an excellent source for solving problems such as the one stated in Example 3.1. It follows from the above that (a) at the pK the concentration of two of the species are equal; (b) moving up or down one unit makes the ratio 1 :10, i.e., there is only 10Y0 of one of the species; (c) movingup or down two units makes the ratio l :100, and one may disconsider the more dilute species in question. Figure 1 is an example of this, phosphate PKa values being 2.12, 7.21, and 12.67. It is a general practice to use polybasic acids (such as phosphoric acid, Fig. 3) as buffers, since they exhibita buffer effect over a considerable pH range. The buffer capacity, 4, is defined as the amount of acid (GAcid) of base (GBase) it would take to change the pH (GpH) by one unit, or on a more differential scale, (3.36) or a similar expression for the base. In kinetic work, there are several factors that are of importance: the concentration of a given species (which for instance in the case of an acid or a base can be calculated from the enders son-Hasselbach equation), its rate constant, and of course the pH. Ong et al. (1964) and Avdeef et al. (1993) have reported that the equation by Yasuda-Shedlovski (Shedlovski, 1956) given below works well. It is akin to the Henderson-Hasselbach equation: (3.37-3.39) Here syA and yHA are the activity coefficients ofprotolytic pair, and pis the fraction converted into more basic species. Fallavena and Schapoval (1997) have employed potentiometric titrations in methanol-water mixtures to determined the PKaof nimesulide and found the equation to work well. They determined the pK of the compound in various concentrations of methanol and to obtain the PKa in water they plot the reciprocal of the dielectric constant of the methanol-water mixture, a so-called Yasuda-Shedlovski plot (Shedlovski, 1956).This is shown in Fig. 1 for nimesulide.

y = 7.6430

+ 0.61'745~

RA2 = 0.995

8.9 r

1OOO/D

Fig. 1 Yasuda-Shedlovsky plot. (Constructedfrom

data published by Fallavena and

Schapoval, 1997, for mimesulide.)

The question of pK values in hydroorganic mixture is of interest, analytically, sincesuch mixtures are often used in liquid chromatography solvents. Barbosa et al. (1997) determinedthe pK values of fluoroquinolonesin acetonitrile and found them to be fairly linear in mole fraction, although pK versus reciprocal dielectric constant also appeared to be linear for norfloxacin, fleroxacin, and flumequine. The apparent pK values of a compound is a function of the dielectric constant of the medium. As an example, Brand1 and Magill (1997) have reported on the pK values of ketorolac in isopropanol. Fig. 2 shows the effect of the medium on the apparent pK. ixtures of Weak and Strong E~ectrolyte§

In calculating the ionic strength for a monobasic acid (for instance acetic acid), the ionic strength is equal to the concentration, C, times the degree of dissociation, a. Since ZA-= 1,Eq. (3.20) becomes

as quoted, e.g., by Maron and Prutton (1965). A practical problem that arises is often that a solution is made with buffer components (e.g., a mixture of NaH2P04 and Na2HP04). This is supposed to give a desired pH (e.g., 7.5), but when the solution is made up, the pH might be only 7.4. It is then adjusted to 7.5 by adding a solution of KOH (see Fig. 3). Although it is possibleto calculate the ion concentrations hypothetically, these calculations become cumbersome, and it is a good practice to burette in KOH (or other base) of known molality, so that its exact volume is known.The pH is known, the concentrations [HA] and [A"] are known from the Henderson-Hasselbach

10 20 30 Percent Isopropanol

ig. 2 PE; of ketorolac versus percent isopropanol. (Figure constructed from data published by Brand1 and Magill, 1997.)

-1

-2 -3

-4

-5 0

1 0 1142

PH Fig. 3 Concentrations of species of phosphate ions as a function of pH. Overall concentration is onemolar. An exampleof an application of this to a drugis that of van der Howen et al. (1994).

equation, the concentration [K'] is known, and [OH"] or [H']is known from the pH. Hence, the ionic strength can easily becalculated. As mentioned, failing to know the exact quantity of base added makes the calculation much more difficult. Once the approximate ionic strength (p1) is known, then a more accurate estimate of the PI((,) for the actual situation can be calculated. This necessitates a new value of a, and in essence the process should be continued until consecutive, iterated values converge. More than one iteration, however, would result in very small changes and is rarely done. L

When a drug A hydrolyzes in aqueous solution, the reaction can be catalyzed by various ionic species, and/or simply be noncatalytic. This is schematized below. The approach is simplified(not distinguishingreaction mechanism, e.g., nucleophilic or electrophilic attack catalysis): Noncatalyzed hydrolysis (rate constant k:)

A + H20 -+ products

(3.4 1)

Specific acid catalysis (rate constant k:) A

+ H' + H20 -+

products

(3.42)

Specific base catalysis (rate constant k?) A

+ OH-

-+ products

(3.43)

HQ (from the HQ/Q--buffer) catalysis (rate constant k & ~ ) A + HQ -+ products

(3.44)

The rate with whichthe reaction takes place is(or may be) a function of the acidity of the medium in which it occurs. If an experiment is to be carried out at a given pH, then this can be achieved either by using a pH stat, and titrate acid or alkaline decomposition products, so as to maintain a constant pH, or a buffer (HQ) may be added. A buffer, by its very nature, may affect the rate, and hence the overall rate of the reaction can be written as

where, for convenience, in this context, the hydronium ion activity is equated with concentration. It is noted that the term [H201 occurs in all the terms, so that it can be merged with the k values to give

As mentioned earlier, the symbol for the hydronium ion has been dropped in favor of [H'], and this will be the convention henceforth. It follows that the k values are 55.5 = 55.5 in dilute solutions). It is noted that this is only times the k* values (since [H201 true for dilute solutions, and a section dealing with concentrated solutions will be presented later. It is seenthat ata given bufferconcentration (and therefore at given hydrogen and hydroxy-ion concentrations, the reaction expression is

dA = --k'[A]

(3.48)

where k' is a pseudo-first-order rate constant given by

Eq. (3.49) predicts a first- or pseudo-first-order reaction, and Eq. (3.46) predicts the rate constant to be a function of severalfactors. It is conventional first to eliminate the effect of the buffer. This isdone by performing the experiments at several buffer concentrations at several pH values (i.e., at several ratios of [HQ] to [Q]). An example of this is the work by Maulding et al. (1975) on the hydrolysis of chlordiazepoxide. Their data are shown in Figs. 4 and 5. Buffer effects will be discussed in a later section of this chapter, and in general the observed rate constants, k&s, are linear in buffer concentration. However, it should be pointed out that they are not always linear over the entire range, as shown e.g. in Fig. 5. Figure 5 shows the hydrolysis rate of chlordiazepoxide in presence of equimolar amounts of phosphoric acid and dihydrogen phosphate ion at 79.5"C, in various buffer concentrations. It is seenthat when the data are back-extrapolated to a zero concentration, the buffer-free rate constant is 0.9 h-l. This extrapolated rate constant will be denoted k in the following. (It should be noted that Maulding et al. found that at phosphate concentrations higher than 0.1 M, linearity was lost,

-0.5

0 ,

-1 .o

-1.5

pH 1975.)

pH profile of chlordiazepoxide. (Graph constructed from data by Maulding et al.,

y L= 0.31286 + 2 , 2 7 1 4 ~ RA2 = 0.999 y = 8,5429e-2 0 . 6 2 5 7 1 ~ RA2 = 0.999

y-0

Fig. 5 Linear buffer concentration plot (panipenem). (Graph constructed from data lished by Ito et al., 199’7.)

pub-

indicating a shift in rate-determiningstep.) In other words, this buffer-free rate constant, k , is given by

k = ko

+k+[H+] (3.50) + k_.[OH-]

The effect of the phosphate-to-phosphoric-acid ratioon the rate constant, in the case o f linearity, can be easily calculated. If one simply considers two ratios (i.e., two pH values), then the equations for the pseudo-first-order rate constants would be (3.51) and

Plots o f k&s at a given pH (not loglo[k]) should therefore be linear in buffer concentration, asshownin Fig. 5. For instanceHou and Poole(1969a,b) showed the traditional buffer effects in the hydrolysis o f ampicillin in solution at different pH values, which enabled them to calculate the catalytic effect o f the different buffer species. Ito et al. (1997) determined the buffer concentration effect on the hydrolysis o f panipenem.

0.6

0.5 0.4

SI 0 v

0.3 0.2 0.1

100 x [E-12P04-1 (~ol~r)

.6 Effect of buffer ratio HQ/Q- of a phosphate buffer on chlordiazepoxide decomposition. (Graph constructed from data published by ~ a u l d i n get al., 1975,) The k values at the two pH values are found from they intercepts at the two pH values, and there are now two equations with two unknowns(kHQand k ~ )so , these can be calculated. Carstensen et al. (1992) have described an abbreviated method for an experimental design that will allow such calculations with a minimum of experimentation. At times the effect is more complicated (Fig. 6), such as showing nicks in the curve, i.e., possessing two line segments. This would implythat the basic equations (3.51) and (3.52) do not apply. This could, for instance, be due to aggregation or micellarization of the drug substance, which might be effected by buffer concentrations. The aggregates may have different degradation rate constants. When the base-ten logarithms of the k values are plotted versus pH, then the so-called kineticpH profiles result. The data from Maulding et al. (1975) are plotted in this fashion in Fig. 2. When the pH is sufficiently low, then

or base-ten logarithmically, Such a plot should be linear in pH with a slope of -1. Figure 4 shows that this is so in the low-pH region. At high pH, only the last term in kobs is important, and (3.55) or base-ten logarithmically, (3.56)

so that a plot of log kobs versus pH should be linear with a slope of +l in the high-pH region. For intermediate pH values, there can be equal weight of the terms k+[H+]and k-[OH-], or the term ko may predominate. In such a case there will be a plateau >k+[H+] and k-[OH"]. where ko > 2.1

Is the

There is a thought that occurs from time to time as to whether, in a simple schemeas above, the optimum pH would equal that of the pKa of the compound. The answer to this question is no, and the failure of the question is exemplifiedfor a situation where HA + H+ -+ products

(3.57)

and A-

+ OH- -+

(3.58)

products

We here have two straight lines as shown in Fig. 7 . The equations for the lines are

(3.60) where at low pH the concentration of HA is approximately equal to the analytically determined concentration of drug, and at high pH, (3.61)

kobs = k-[OH-][A"]

-4 0

7' A V-shaped, achematic pH profile. An example of this is the work by King et al. (1992). In reality, there is a gradual transition at the minimum, not a sharp break.

where at low pH the concentration of A- is approximately equal to the analytically d e t e ~ i n e concentration d of drug (see Fig. 7). The intersection of the two lines with Eqs. (3.60) and (3.61) occurs when (3.63)

k+[H+][HA] k-[OH"][A-] Introducing Kw and K(,) now gives (3.64) or [H+] == 1 0 - ~

(3.65)

and this is not solely a function of If, indeed, the maximum stability occurs at pH =PK(a), then this would be quite fortuitous. There are many different types of pH profiles, and some of these will be discussed in detail in the following. erat~reEffects

At each given pH, each of the rate constants (ki)will adhere to the Arrhenius equation (Fig. 8). If there is a domain where one rate expression (kq)is predominant, then

where (b(q) can be unity or [H+] or [OH-]. In such regions the Arrhenius equation can beexpected to hold, but the activation energies for the various kq terms may differ. In regions of pH where more than one term is predominant then, denoting these by kql and kq2 it follows that kobs = kql#(ql) (3.67)

+kq2#(q2)

so that (3.68) and unless

is the same over the region studied, a linear Arrhenius plot is not to be expected.

2.8

2.9

3.1

3.0

1ooorr

3.2

ig. 8 Arrhenius plot of aqueous decomposition of chlordiazepoxide at pH 11.The least-squares fit of the line is ln[k] ==25.05-9.61 (10001T)(R= 1-00) showing an activation energy of 19.2 kcallmol. This may be compared with the activation energy of 20 kCal/mol used in the concept of KMT. (Figure constructed from data published by Madding et al., 1975.)

Pl-1

Effect of temperature on optimum pH.In the figure, theactivation energies are such that the valueof kt.at 40°C is1.58 times that at25”C, and for k- the valueat 40°C is7.9 times that at 25°C. The pH shift is quite obvious.

Furthermore, it is not to be expected that the “optimum pH” will be the same at all temperatures. This is demonstrated in Fig. 9. talysis

It should be pointed out that when reporting kinetic pH profiles, it is most instructive to drop out the buffer effectterms. This isoften not done, at the loss of some insight into the problems at hand. There are many examples ofpH profiles, e.g.,Ravin et al. (1964) reported on kinetic pH profiles of carbuterol.

For pHprofiles in general, obviously there are always, in buffer systems, competing reactions, since there is general acid or base catalysis (buffer catalysis) or specific acidor base catalysis ([H+l or [OH-]). If k simply denotes the bimolecular rate constant for the particular pathway (i.e., either general or specific), and if p is the number of nonequivalent protons, dissociable on the conjugate acid bade, and q the amount of sites of the general base that can accept a proton, then the relation (3.70)

is denoted the Brransted relation. p is the Brmnsted coefficient and GB is a constant. A s mentioned, pH, is the negative l o g a r i t h ~of the conjugate acid dissociation constant (and for hydroxide attack, for instance, is pKw -pH). ost treatises deal with acid catalysis. Kirsch and Notari (1984) have dealt with the theoretical basis for the detection of general base catalysis in the presence of predominating hydroxide catalysis. This is difficult for two reasons. Firstly, the pH range were the stability is the least is also the range where the specific catalysts (i.e., [OH]) predominate over buffer catalysts. Most often buffer catalysis is completely overlooked in this region. As a general statement, it can seen from the above that when the degradation of a drug accelerated is by general and specific base catalysis, then the pseudo-first-order rate constant for loss of drug, kobs, is given by (3.71)

where ktH is the rate constant (bimolecular) for specific base attack at zero ionic strength, kg is the rate constant (bimolecular) for general baseattack, and the terms in square brackets are hydroxyl and specific base concentrations. The f terms are activity coefficients, and 3! denotes transition state. General base cannot be distinguished from specific hydroxideand general acid catalyses. The latter is defined as (3.72)

where Ka and Kw denote buffer and water dissociation constants, and where HB denotes the acid conjugate to the base. If this is divided by kobs then the fraction, P ,due to general basereaction is obtained. ~ntroducingthe Brransted relation into gb this then gives (3.73)

This term as a function of the pK, valuesof a series of buffers has been calculated by Kirsch and Notari (1984). When the Bransted coefficientisclose to unity, the

equation simplifies to (3.74) Kirsch and Notari (1984) testedthe equation using the conversion incarbonate buffers at 60°C and unit ionic strength of ancitabine prodrug. The study was carried out at unit ionic strength. Their data showed excellent correlation between observed values of Fgb and those calculated.

In the following, an acid willbe denoted HA and a base B. In the former case there is a possibility that HA decomposes and A- does not, and the possibility that both decompose. (H2A+ and A= are also prone to decomposition in certain cases as will be seen in the following). A similar situation exists with a base, B, where both BH+ and B (and in cases BH;+ and B-) maydecompose, others whereonly onedecomposes. There is,basically no differencebetween the HA and the B terminology, but it is maintained in the following for convenience. Equations derived for one will applyto the other as well. In the immediately followingtext the species HA will be considered. If an acid is only hydrogen-ion catalyzed, then

HA + H+

+ (H20)

products

(3.75)

products

(3.76)

" +

and A-

+ H+ + (H20)

" +

The quantity monitored, analytically, is the drug concentration, C, given by (3.77) C = [HA]

+ [A"]

where [HA]and [A-] are related to one another by the acid equilibrium constant K(a), of the compound: (3.78) or (3.79) From this it follows that [A-] = C -[HA] = C -[A-I[H+l

(3.80)

~ a r s t ~ ~ s ~

from which (3.81) and (3.82) These expressions are introduced into the rate equation:

(3.83)

Hence the reaction is apparent first order with an observed rate constant, kobs, of (3.84) when plotted in this fashion, curvature will be followed withthe appropriate fitting of parameter values (/cl+, k2+, and Ka). The latter is usually known,independently. The last term in Eq. (3.84) can be written (3.85) At low [H+] this term vanishes, and the first term becomes (3.86)

(3.87) as expected, and a similar argument willshow becomes

The logarithmic forms of Eqs.(3.87)

and

thatat

high[H3-]expression

and (3.88) are

so that a plot oflog[k2+] versus pH willbe twolinesegments(with different intercepts) both with a slope of minus one. This is the type profile which later (inFig.8)is denoted type JDEN. A similar situation where [HA] is only hydroxylion catalyzed will give profile of type AHJK (Fig. 11). Of course, when more species are involved, then the kinetic expressions can become quite involved (e.g., the work on decomposition of alprazolam by Cho et al., 1983) and frequently include several terms of the type (3.9 1)

Another instance is the specific acid/base catalysis of a compound present as only one species. The reaction stated would be HA.

+ H+ + (H20) -+ products

(3.92)

+ OH- + (H20) -+products

(3.93)

and

A s shown in Eq. (3.82),

[HA] ==: C -[A-] = C

[H+] EH+] K(a)

(3.94)

The rate equation is

"dC -k+[HA][H+](3.95) +k_[HA][OH-] dt

Introducing Eq. (3.82) and the expression [OH"] =Kw/[H+] intoEq. (3.95) gives (3.96) i.e., it is first order in C with an apparent rate constant of (3.97) where Kw is water's ionization constant. At high pH values, [H'] < /[H+] is large and [H+] 1s not large. The last term then is

(3.98)

~ar§ten§en

(3.99) i.e., is constant. At low pH (high [H']) the last term in Eq. (3.96) vanishes and the equation becomes (3.100) or

and at high pH (low [H']) the first term in Eq. (3.96) vanishes and it becomes (3.102) i.e., is at a plateau, since the right-hand side is a constant. At intermediate pH values the last term will predominate, i.e., (3.103)

i.e., the pH profile will have slope -1 at low plc, plus slope 1 at intermediate slope, and be horizontal at high pH. This corresponds to a profile denoted ABCDE at a later point (Fig. l l ) , An exampleof this is pralidoxine (Ellin, 1958; Ellinet al., 1964; Ellinand Wills, 1964).

Most decompositions in solution are, as mentioned, hydrolyzes. If the drug substance is an acid, then in the final product the pH will be fixed. The situation is complicated by the number of species in solution and by the fact that the pH is usually controlled (kept constant) by the addition. of buffers. The effect of pH will be discussedin the subsequent section; suffice it tosay here that in cases whereboth undissociated acid (HA) and anion (A-) are fairly soluble, the following holds: A+H20 +E

(3.105)

Kinetic pH Profiles

with rate constant k, and (3.106) A"

+ H20 + E'

with rate constant k'. E and E' stand, globally, for decomposition product. The quantity measured analytically is usually the sum, C, of the concentrations of free and ionized acid, i.e.,

(3.107) C = [HA] [A"] The rate equations for disappearance of overall concentration, C, remaining at time t , is given by dt

= -k[HA] -k'[A-]

(3.108)

The equilibrium for the acid is HA-A-

+H+

(3.109)

The equilibrium constant is denoted K&, so A" = K(a)[HA][Hs]

(3.1 10)

Equations (3.107), (3.108), and (3.1lo), if the pH is kept constant, contain four unknowns (including dCldt) and can therefore be reduced to one equation with two unknowns. This reduction can be done in such a fashion that the unknowns are C and dCldt, and so a differential equation results, which allows solvingfor C. Inserting Eq. (3.110) into Eq. (3.107) gives (3.1 11) or [HA] =

C 1 K/[H+]

(3.1 12)

Inserting Eq. (3.111) into (3.107) then gives

C [A-] = C 1 K/[H+]

(3.113)

Equations (3.1 11) and (3.1 13)are now inserted into Eq. (3.108) to give (3.1 14) which may be rearranged to read (3.115)

~~rste~sen

where (3.1 16) Equation (3.115) integrates to (3.1 17) which of course is first order. Changing the pH will have the effect shown in Eq. (3.116). LlFl

FIL

It stands to reason that the stability of a drug substance in solution depends on its molecular environment. One of the most important macroscopic parameters of this nature is the pH, and a large volume of literature exists on this subject, in particular on the effect of pH in the acid range. A simplified approach is presented first. In the case of hydrolysis of a nondissociating drug the decomposition is given by A+H20+ E

18)

(3.1

where E, again, is a general symbol for decomposition products. The rate of the reaction is (or may be) a function of the acidity of the medium in which it occurs. If an experiment is carried out at a given pH, then this can be achieved by using a pH stat, or a buffer (HB) may be added. A buffer, by its very nature, can affect the rate, and hence the overall rate of the reaction can be written d[A1 --[A] {ko dt

+k+[H"] +k-[OH"] +~HB[HB] +k~[B-l} [H201 (3.1 19)

where, for convenience, in this context, the hydrogen ion activity is equated with concentration. The subscripts to the rate constants, k, in Eq. (3.119) refer to, in the order given, the hydrolysis per se, the hydrogen ion catalyzed, the hydroxyl ion catalyzed, the buffer acid catalyzed, and the buffer anion catalyzed rate constants. It is seenthat ata given buffer concentration (and therefore at a given hydrogen and hydroxy-ion concentration) the reaction expression is dA dt

= k[A]

(3.120)

where k is a pseudo-first-order rate constant given by

Equation (3.120) predicts a first- or pseudo-first-order reaction, and Eq. (3.121) predicts the rate constant to be a function of several factors. It is conventional first to eliminate the effect of the buffer. This is done by performing the experiments at several buffer concentrations at several pH values (i.e., at several ratios of [HB]

to [B]). An example of this is the work by Maulding et al. (1975) dealing with the hydrolysis of chlordiazeopxide cited earlier. Their data are shown in Fig. 4.

he most common approach is the use of buffers HBI where HB denotes an acid. Since the buffer may, as seen above, affectthe kinetics, it is customary to carry out experiments where the buffer concentration is varied. As shown in Fig. 10, this type ofplot is usuallylinear (at low concentration of buffers), and the procedure generally used isto extrapolate the line to zero concentration to obtain the "buffer-free" rate constant, k*. This is done at two pH values (pH1 and pH2) obtained by the same buffer system. The k* values are found from the intercepts of the straight lines at the two pH values, and the slopes yield two equations with two unknowns (khb and kb), so that these can be calculated. When the Briggsian (i.e. base-ten) logarithms of the k* values are plotted versus pH, then the so-called kinetic pH profiles result. The data from Maulding et al. (1975) are plotted in this fashion in Fig. 4. kobs, at given pH values, may or may not follow an Arrhenius equation. A case where theydo so is shownin Fig. 8. There are many different types of pH profiles. Ashas been seen in some cases, when the pH is sufficiently low, then k* m k+[H+]

(3.122)

or, logarithmically, lo~,o[k*l= (3.123) log10[k+l -PH Such a plot should be linear in pH with a slope of -1.

0.06 0.04 0.02

Acetate Concentration, M Graph constructed fromdata published by Fassbergand Stella (1992) dealing with the hydrolysis of campothectin. Similar examples are those of Maulding et al.(1975) and of Ito et al. (1997).

~~rstensen

At high pH, only the last term in k* is of importance, and (3.124)

(3.125) i.e., a plot of log k* versuspH should be linear with a slope of +l. An exampleof this is the hydrolysis of 3-quinuclidinyl benzilate indilute aqueous solution reported by Hull et al. (1979). For pH values in between there can be an equal weight of the two terms k+[H+] and k-[OH-[, or, when there is a plateau, the term ko may dominate. It should be pointed out that when reporting kinetic pH profiles, it is most instructive to drop out the buffer effect terms. This is often not done in practice, at the loss of some insight into the problems at hand. There are many examples of the above, for instance the kinetic pH profiles of carbuterol reported on by Ravin et al. (1964). Hou and Poole (1969a,b) showed the traditional buffer effects in the hydrolysis of ampicillin in solution at different pH values, which enabled them to calculate the catalytic effect ofthe different buffer species. S~ECIFIC ACID

For pHprofiles in general, there are always, in bufferisystems, competingreactions, since there is general acidor base catalysis (buffer catalysis) or specific acidor base catalysis ([H+] or [OH-]). In this case, denoting by HA and A the nonionic and ionic form of the drug substance, presuming it to be an acid, the acid equilibrium would be (omitting, for convenience, water and using hydrogen rather than hydronium ion, and using concentrations):

HA+-+H+ + A -

(3,126)

with equilibrium constant (3.127) The two species hydrolyze with pseudo-first-order rate constants kHA and kA, i.e., using the notation D = d/dt, L)[HA]

-~HA[HA]

(3.128)

and D[A-] = -k~[A-l

(3.129)

The analytically measured concentration is C, given by C = [HA]

+ [A")

(3.130)

Rearranging Eq. (3.12'7) gives (3.131) Introducing this into Eq. (3.130) gives (3.132) hence (3.133) The sum of Eqs. (3.128) and (3.133) is DC = -~HA[HA] -k ~ [ A " ] z- ~ H A [ H A] kA{C -[HA]}

(3.134)

DC = (kHA -~A)[HA] -kAC

(3.135)

where Eq. (3.130) has been usedfor the next-to-last step. Introducing Eq. (3.134) into Eq. (3,135) now gives kA C --(!.c=-

_ . _

dC dt

(3.136)

where (3.137) The above is the simplest of all treatments of the situation described in the heading. A good example of this is the work by Longhi et al. (1994). If the rate constants kHA and kA are functions of [H+], [OH"], and buffer concentration, then the expressions becomesquite complicated. Often some of these are of importance, others not, and inthesecases it there willalwaysbe one term of the type k/{ 1 K/[H +l>. For instance for the decomposition of moxalactam Hashimoto et al. (1984, 1985) arrived at the equation

(3.138) where ko and kd represent the first-order rate constants for the water catalyzed degradation of dianionic and trianionic moxalactam. This curve will give the usual -1 slope in the acid (in this case below pH 2) region, and a +l slope in the alkaline (in this case above pH 8) region.

When the processes becomemore complicated, then the kinetic expressionscan become quite involved (e.g.,the work on decomposition"ofalprazolam by Cho et al., 1983). . T

To arrive at an expression for the "buffer-free" observed rate constant as a function of pH, the general procedure is as follows: 1. To write up all the species in question. These in general consist of buffer (HB) and drug (HA) species. 2. To write up all the equilibrium constants existing between various ionic forms of the same species. 3. To express the analytical concentration in the form of the species present. 4. To write up all possible reactions (HA-product, HA+H+ "+product, HA OH- "+product, and so on). pertinent. It will be seenlater that 5. To eliminate those reactions that are not in the absence of a kinetic salt effect, reactions of two charged species (e.g. A"+H+ -+product) are unlikely. 6. To write the rate constants for the various drug species. 7. This should produce n equations with n 1unknowns. 1ldt is usually denoted 1)and treated as a constant, and the equations are solved for each species. These can be reduced in traditional fashion to one equation with two unknowns (dCldt and c>, and this is the differential equation that is sought.

When the equation for the measured species as a function of pH is arrived at (and it can be quite complicated), then there willbe M rate constants, and M + K data points (i.e., K degrees of freedom). The data are then fitted by nonlinear programs in a computer, or curves are generated with different values of the M rate constants until one set of values is found that fits the data.

These curves a r e ~ r e q u e ~ tcalled ly "theoretical"curves, but they are indeed but best fits. It is quite possible that there is some reaction with a rate constant that

is not taken into account, or conversely that there is one more reaction accounted for than necessary. The shape of the profile can vary and will have the shape of one of the line segments in Fig. 11. For instance the profile shown in Fig. 4 would be of the type ABCDE. The various schemes associated with di€ferent profile shapes will be discussed in the following. Van der Houwen et al. (1988, 1991 ,1994, 1997) have analyzed the general procedure and called to our attention flaws in many interpretations on the subject. They have summarized the possible situations by considering the situations where 0, 1, 2, 3, and 4 protolytic equilibria are at hand. In their treatise as well as here, the subscripts 0, 1,2, etc. are used as terminology for the successive deprotonation, e.g., subscript 0 is used for the fully protonated species. For an amine,RNH' t, RH H+ would have an equilibrium constant of KO and superscripts H, S, and OH are used for hydrogen ion, solvent, and hydroxyl ion catalysis, so that for the fully charged amine referred to its rate constant with

hydrogen ion would be designated ky .Van der Howen et al. (1997)take into account kinetically indistinguishablesituations, and for this purpose they introduce the terms (3.139) (3.140) (3.141) (3.142) (3.143) (3.144)

= k~~KwKalKa2Ka~

(3.145)

and the applicable equations are listed in Scheme I.

kobs = MO[H'M1 ]

[H+]

(3,146)

(3.147)

(3.148)

In preliminary stability work, the existing problem isusually shortage of drug substance. Another problem that often plagues the preformulation phamacist/chemist is that the assay maynot be sufficiently developedto, e.g., label all the decompositionproducts. If an assay isHPLC, a decomposition product could hide under the parent peak, and kinetic plots still appear first order. (TLC should always be resorted to guard against this possibility.) A pH profile accounting for buffer effect (i.e., extrapolating to zero buffer concentration) can be done, even with smallamounts of drug, but kinetic salt effect and the finer points of reaction mechanisms cannot. It suffices, in any event,to establish the best buffer, and the optimum pH range, and to produce a fairly good pH profile. If the terminal slopes are +l and -1, then of course the simple acid and base catalyses have been established in those regions. The pseudo-first-order of the reaction has also been established allowing reasonable extrapolations, and the types of conclusions needed for stability work. It will also aid in the setting of expiration periods. A s work iscompleted, it is goodto repeat the mechanistic work in more detail, and to elucidate mechanisms. After a drug is introduced into the market place, at most one year should be allowed to pass before a detailed pH profile is known(and preferably submitted for publication). The drug in the following will be symbolized as HA. (or, if it is a zwitterion, HA*), its anion A-, and its positively charged species HzA+. An alternate symbol presentation also used (usually whenthe compound is, in general, considered a base) is to denote it B and to denote the ionized species HB’. The patterns in Fig. 11 are drawn as line segments,but of course, in reality the “corners” are smooth. But for schematizing purposes, the lines in Fig. 11 will suffice. In areas where oneterm predominates in the kobs expression, this is quite acceptable, but when twoterms become predominant, then it oversimplifies the actual profile. In te:rnpting more exacting expressions for these situations, the complication is that

b l

”

1 Generalized pH profilepolygon.

the analytically observed species,C, is not equal to either [A-] or [HA], for instance, but is the sum of these two terms.

.l. This type (Fig. 12)would most often be attributed to hydrogen-ion catalysis, i.e., in the simplest case, to [HA]

+ [H']

-+ products

(3.151)

It is usually, when reported, part of a study over a rathernarrow pH range and hence could be but part of a more complicated scheme. The necessityfor only studying a compound ina narrow pH range can be dictated by many factors, such as solubility, for instance, Chloramphenicol (Higuchi et al., 1954) is of type AB and is studied from pH 0.7 to 1.7. Digoxin (Khalil and El-Masry, 1978)is Type AB and is studied from pH 1 to 2.2. This type has also been reported for tri~uoromethoxyphenyl cyanopropionamide byBrand1 and Kennedy (1964) and is cited elsewhere. Hydrocortisone sodium phosphate isof this type (Marcus, 1960) and has been studied from pH 6 to 7.5. In this latter case the slope is not -1, and the rate is assumed to be (3.152)

Rifampin (Seydel, 1970)is of type AB and again was studied only in a narrow pH range (0.5-1.5). Polyortho esters of 3,9-dibenzyloxy-3,9-diethyl-2,4,8,l0-tetroxaspirol[3,5]undecane (Nguyen et al., 1984) are of type AB in a pH range of 3-8. Longhi et al. (1994) have reported the pH profileof the subtype AB for isoxazolylnaphtoquinone.There is no ionic strength effect, so the equation, as seen, becomes simply k =k[H+].

pH (Graph constructed from data from Longhi et al. (1994).) The least squares fit is

kobs

= 0.23 -(l.OpH).

g -2.8 -e T-

“3.0

rJ)

-3.2 -3.4

c -3.6 0

PH 3 ka from dalvastin kinetics. A profile of type ABC. (Graph constructed from data published by Won, 1994.)

Type AB may simply be a part of more complicated profiles, had the study, extended over a larger pH range.

A profile of typeABC (Fig. 13)isexemplified by the workpublished on epimerization and hydrolysis of dalvastin by Won (1994). Lactone formation of the HMG-CoA reductase inhibitor, (3-981, which is l-methyl-ethyl)-3-phenyl-4-[(phe[R-(R*,R*)]-~-(4-fluorophenyl)-~,~-dihydro~y-5-( nylamino)~arbonyl]-lH-pyrrole- l -heptanoic acid followsthis scheme. The hydrolysis of this compound has been reported by Kearney et al. (1993), and isof an ABC type, with a break at pH 4.5. This profile can often be attributed tohydrogen-ion catalysis and noncatalysis of the hydrolysis of HA, and has an observed rate constant of

w h e r e f j ~is the fraction present as HA at the pH in question. This, for instance, holds for furosemide (Cruz et al., 1979). Promethanzine (Underberg, 1978a,b; Stavchanski et al., 1983), pH 1.5-5.5, is of this type with slope 1. This is complicated by the fact that the compound forms micelles above its critical micelle concentration.

This type (Fig. M), in general, isa profile where onlyhydro~enionand hydroxylion catalysis of one drug species plays a part. If there are horzontal parts then this is often attributable to HA+H20 ”+products. This is, for instance, the case in

-6

Subtype ABCL). kb from dalvastin kinetics. A profile of type ABC. (Graph constructed from data published by Won, 1994.)

meperidine and methylphenidate below. Drug substances reported to be of type Acetaminophen (Koshi and Lach, 1961), pH 2-9 9-~-Arabinofuranoxyladenine acyl migration (Anderson et al., 1985) (Hamilton-Miller, 1973), pH 4-8 Atropine (Lund and Waaler, 1968; Zvirblis et al., 1956), pH 0-10 5-Azacytidine (Notari and DeYoung 1975), pH 4-9 Benzylpenicillin (Brodersen, 1947; Finholt, et al. 1965), pH 1-11 erge et al., 1983; Yamana and Tsuji, 1976), pH 0-10 Cephalothin (Yamana and Tsuji, 1976), pH 0-1 1 Cefixime (Nakami et al., 1987) Chlorothiazide (Yamana et al., 1968), pH 3-11 Codeine sulfate (Powell, 1986), pH 0-12 Cyanocobalamine (Loy, 1952), pH 0-9 Cyclophosphamide (Hirata et al., 1967), pH 0-14 Dalvastin (Won, C. M. 1994), pH 1-10 Diazepam (Han et al., 1977), pH 0-12 Ergotamine (Kreldgaard and Kisbye, 1974), pH 2-4 d Takla, 1968), pH 5-10 Erythromycin (Am R*)]-2-(4-fluorophenyl)-~,~-dihydroxy-5-( 1-methylenylamino)carbonyl]-1H-pyrrole-1"heptanoic acid (Kearney et al., 1993) break at pH 4.5. pH 1-9 ethylphenidate (Siege1 et al., 1959) pH 1-6

arst~nsen

Meperidine (Pate1 et al., 1968), pH 2-7 Methylphenidate (Siege1 et al., 1959), pH 1-6 Succinyl choline chloride (Suzuki and Tanimura, 1967), pH 1-7 Sulfacetamide (Meakin et al., 1971), pH 0-13 Atropine (Lund and Waaler, 1968;Zvirblis et al.,1956) atpH below 4 attributed the decomposition to hydrogen ion catalysis, and above 6 it follows the equation rate = k-{[OH"][B]}

+k'[OH-][HB+] (3.154)

In the case of phenylbutazone (Stella and Pipkin, 1976) the shape is ABCD,but the slopes are not +l or -1. In this case there is an enolequilibrium (withequilibrium constant K*)

(3.155)

Oxazepam (Han et al., 1977) isalso of type ABCD, but with nonunity slopes. In this case the molecule is considered a diprotic acid (HzA), and the fractions of diprotic, monoprotic, and unchargedspecies are denoted f2, f i , and f o , respectively. The value for kobs is then

i.e., a combination of hydrogen or hydroxyl ion catalyzed reaction of the three protic species. Indomethacin (Pawelczyk et al., 1979) is another case where subtype ABCD applies but where the slopes are nonunity, and this is accounted for by

Methicillin (Schwartz et al., 1965; Hou and Poole, 1969a,b) also shows nonunityslopes.Penethicillin(Schwartz et al., 1962) also is accounted for by this equation. Scopolamine (Smithuis, 1969; Guven and Aras, 1970) and pilocarpine (Baeschlin and Etter, 1969; Chung et al., 1970) are of type ABCD but more cornplicated than the remainder of the reactions quoted. King et al. (1992) had reported the pH degradation profile of moricizine to der Houwenet al. (1988,1991,1993,1994,1997) be of typeABCD.Van reanalyzed the data and concluded thatit wasof type ABCDE. Fubara and Notari (1998) have shown that the degradation of cefepime in aqueous solution isof type ABCD.

inetic pH Profiles

This is type ABCD with a horizontal part at the end (Fig. 15). The simplest mechanism that this suggests is HA

+ H'

-+ products

(3.157)

HA -+ products

(3.158)

A -+ products

(3.159)

+OH-

(3.160)

-+ products

Aspirin (Garrett, 1957) has a pH profile explained by this scheme. Other mechanisms accounting for the profile exist: in the case of pralidoxine (Ellin and Wills, 1964),pH 0-14, it is attributed to hydrogen and hydroxyl ion catalysis of P+ giving rise to the equation (3.161)

>K a , At low pH where [H+[ > (3.162) giving rise to the minus unity slope of log k versus pH. At intermediate pH where

5.3. Polymorphism

245 247 248 248

6. Vapor Pressure

249

7. Partition Coefficient

25 1

8. Hygroscopicity

25 1

9. Compatibility Tests 9.1.Useof DSC 9.2.Use of microcalorimetry 9.3. Compatibility test for solid dosage forms 9.4. Compatibility with containers

252 253 253 254 256

x., A

10. Kinetic pH Profiles

256

11.Liquid ~ompatibilities 11 1. Aqueous solution compatibility 11.2. Nonaqueous liquids 11.3.Emulsions 11.4. Gels

257 257 257 258 259

References

259

Historically, preformulation evolved in the late 1950sand early 1960sas a result of a shift in emphasis in industrial pharmaceutical product development. Up until the mid-l950s, the g neral emphasis in product developmentwas to development elegantdosage forms, and organoleptic considerations far outweighedsuch (as yet unheard of)considerations as whether a dye usedin the preparation might interfere with stability or with bioavailability. In fact, pharmacokinetics and biopharmaceutics were in their infancy, and although stability was a serious consideration, most analytical methodology was such that even gross ,decomposition often went undetected. It was, in fact, improvement in analytical methods that spurred the first programs that might bear the name “preformulation.” Stability-indicating methods would reveal instabilities not previously known, and reformulation of a product would be necessary. When faced with the problem of attempting to sort out the component of incompatibility in a 10-componentproduct, one might use many labor hours. In developing new products, therefore, it would be logicalto check, ahead of time,which inco~patibilitiesthe drug exhibited (testing it against common excipients). This way the disaster could be prevented in advance. A further cause for the birth of preformulation was the synthetic organic programs started in manycompanies in the 1950s and 1960s. Pharmacological screens would showcompounds to be promising, and pharmacists were faced with the task of rapid formulation. Hence they needed a fast screen (i.e.,a preformulation program) to enable them to formulate intelligently. The latter adverb implies that some of the physical chemistryhad to be known, and this necessitated determination of physicochemical properties, a fact that is also part of preformulation.

The approach of preformulation was so logical, indeed, that it eventually became part of the official requirements for INDs and NDAs (Schultz, 1984): New drug substancesin Phase I submission. For the drug substance,the requirement includes a description of its physical, chemical or biological characteristics. We in the reviewing divisionsregard stability as oneof those characteristics. The requirement of NDA submissions ... of the rewrite stability information is required for both the drug substance and drug product. A good time to start toa c c u ~ u l ainformation t~ about the appropriate methodology and storagestations for use in dosage form stations for use in dosage form stability studies, therefore, is with the unformulated drug substance.. ..Stress storage conditions of light, heat and humidity are usually used for these early studies, so that the labile structures in the molecule can be quickly

identified. ...If degradation occurs, the chemical reaction kinetics of the degradation should be determined.. ..Physical changes such as changes from one polymorph to another polymorph should be examined.. .,With the drug substance stability profile thus completed, the information should be submitted in the IND submission.

The goals of the program are therefore (1) to establish the necessary physicochemical parameters of a new drug substance, (2) to determine its kinetic rate profile, (3) to establish its physical characteristics, and (4) to establish its compatibility with common excipients. To view these in their correct perspective, it is worthwhileto consider when, in an overall industrial program, preformulation takes place. The following events take place between the birth of a new drug substance and its eventual marketing (it is a fact, however, that most investigational drug substances never make itto the marketplace for one reason or another): 1. The drug is synthesized and tested in a pharmacological screen., 2. The drug is found sufficiently interesting to warrant further study. 3. Sufficient quantity is synthesizedto (a) perform initial toxicity studies, (b) do initial analytical work, and (c) do initial preformulation. 4. Once past initial toxicity, phaseI (clinical pharmacology) begins and there is a need for actual formulations (although the dose level may not yet be determined). 5. Phase I1 and TI1 clinical testing then follows, and during this phase (preferably phase 11) an order of magnitude formula is finalized. 6. After completion of the 7 - After approval of the N

Physicochemical studies are usually associated with great precision and accuracy, and in the case of a new drug substance would include studies of (a) p the drug substance is an acid or base),(6) solubility, (c)meltingpoi polymorphism, (d)vapor pressure (enthalpy of vaporization), (e) surface characteristics (surface area, particle shape, pore volume), and (f) hygroscopicity. Unlike in the usual physicochemical studies, an abundance of material is usually not at hand for the first preformulation studies: in fact, at the time this function starts, precious little material is supplied, and therefore the formulator will often settle for good estimates rather than attempt to generate results with four significant figures. There is another good reason not to aim too high in the physicochemical studies of the first sample of drug substance. In most cases the synthesis is only a first scheme, and in later scale-up it will be refined; and in general the first small samples contain some small amount f impurities, which may influence the precision of the determined constants. utit isnecessary to know, grosso ~ 0 ~ important properties such as solubility, p ,and stability. These are dealt with in order below.

0 ,

nse

The definitionof pK(,) and pk, have been discussed in chapter 2. For substances that are carboxylic acids (HA) it is advantageous to determine the p , since this property is of importance in a series of considerations. For carboxylic acid the species A- usually absorbs in the ultraviolet (UV) region,and its concentration can be determined spectrophotometrically (Underberg and Lingeman, 1983); HA on the other hand will absorb at a different wavelength. The molar absorbances of the two species at a given wavelengthare denoted EO and E- (it is assumedthat at the wavelength chosenEOc E-), and it can be shown that if the solution is m0 Molar in total A, then

so that the ratio A-/HA can be determined in a series of buffers of different pH. ence the pK(,) can be found as the intercept by plotting p as a function of log[(A-) /(HA)] by Henderson-Hasselbach:

If several buffer concentrations are used, extrapolation can be carried out to zero ionic strength, and the pK, can be determined. For initial studies, however, a a) in the correct range (i.e., +0.2 unit) will suffice, so that the determination e can be done at one buffer concentration only. The conventional approach is to do titrations (Fig. l), and this will yieldgraphs Usually, the water is titrated as well of fraction (neutralized (x) as a function of rke and Davis, 1954), and whatis pre ed in Fig. 1 is the “different.” The is then the pH at half neutralization (whichis also the inflection point).

Acid

8 I

6 4 2 -

b l

Moles Base Added/ Typical titration curves. The “water” curve indicates the amount of alkali needed to “titrate” the water, and the “acid” curve isa conventional titration curve. The difference curve is the horizontal difference between the “acid” and the “water” curve and is the adjusted titration curve. For example the point “b’’is “c” minus “a”. The is the point of inflection, which is also the point where half of the acid is neutralized.

solubility curvecan now beconstructed sim (at low pH) and A" (e.g., of NaA) at hig

important goal of the preformulation effort is to devise a method for making tions of the drug. Frequently, the drug is not suf~cientlysoluble in water itself to allow for the desired concentrations, for example for injection solutions. ~olubilitiesare detemined by exposing an excess of solid to the liquid in question, and assaying after ~quilibriumhas been established. This usually is in the range 60 to 72 h, and to establish that e~uilibrium indeed has been established, sampling at earlier points is necessary. Unstable solutions pose a problem in this respect and will be dealt with in more detail later. Solubilities cannot be determined by precipitative methods (e.g., by solubilizing an acid in alkali and then lowerin desired pH) because of the so-called metastable (solubility) zone rnedo, 1984). In the writing to follow, drug substances are su categories: (1) ionizable substances, and (2) (virtually) nonioniz stances. ~olubilitydeterminations are necessary both for stability reasons and for formulation reasons. It was noted, in the chapter dealing with stability of solids in the ceof water, that the solubility term becomes part of a rate constant. Since Pre ation occurs in the early stage of development, the optimization of stability by way of compound selection (correct salt) is of importance, and often a drug product can be stabilized by keeping the solubility of the drug substance low. In it this, however, might affect bioavailability. robably among the most well-known examples of suchstabilization are that of procaine penicillin and that of potassium clavulanate. In the latter case, the sodium salt, for instance, is unstable to such an extent that it cannot be utilized. The decrease in solubility of the potassium salt renders the product machinable (although low humidities must be observed in manufacturi~g). cr

amount in solution in a solubility experiment is

where S denotes solubility. The last term can be determined from knowledge of the ,and the use of Eq. (2). gs that areamines, the free base is frequently poorly soluble, and in this g the titration is a solvent containing is often estimated by this at different organic solvent consolvent (e.g., ethano ation can be carried out to 0% solvent centrations (e.g., S%, lo%, 15%, 20 concentration to estimate the aqueous PIC(,). Usually, alkali metal salts of acids are more soluble than the free acids,and in the case of basic (e.g., amine type)drugs, the solubility of the acid addition salts are more soluble than th ases. At times(e.g., in the case of enalipril) the compound is amphoteric. id addition salt issoluble, the freebaseislesssoluble, and the sodium salt is, again, more soluble. In simple cases, the solubility curve *

-8

1 0 1 2

PH pH solubility plot of imiquimod in water. (Graph constructed from data published by Chollet et al., 1999.)

willsimplymimic the Henderson-Hasselbach equation. Anexample of thisis imiquimod (Chollet et al., 1999), the pH solubility profile of which is shown in Fig. 2. Streng and Yu (1998) have published a computer program for prediction of stability curves based on the pK, value of the compound in question.

For hydrophobic, (virtually) nonionizable substances (i.e., those that show no ionic species of significance in the pH range 1 to 10, e.g., diazepam), solubility can usually be improved by addition of nonpolar solvents. Asidefrom solubility, stability is also affected by solvents either in a favorable or in an unfavorable direction ( Niazi, 1983). Theoretical equations for solubility inwater (Yalkowsky and Valvani, 1983) and in binary solvents(Acree and Rytting, 1983)havebeen reported in the literature, but in general the approach in preformulation is pseudoempirical. ost often the solubility changes as the concentration of nonpolar solvent, C2, increases. For binary system, it may simply be a monotonical increasing function (Carstensen et al., 1971), as shown in Fig. 3. The solubility is usually tied to the dielectric constant, and in a case suchas that shown in curve A, the solubility is often log-linear when plotted as a function of inverse dielectric constant, E, that is,

+ e2

l n S = - el ~

where E is the dielectric constant and the e terms are constants (Underberg and Lingeman,.1983).

.3 (A) Solubilityof 7-chloro-l,3-hydro-5-phenyl-2H1,4-be~odia~epine-~-one-4-oxide in aqueous propylene glycol. (Data from Carstensen et al., 1971.) (B) Solubility of another benzodiazepine. (~npublisheddata.) Frequently, however, the solubility curvehas a maximum (as shown in curve in Fig. 3) when plotted as a function of C2 and E (Paruta and Irani, 1964). In either case it is possible to optimize solubility by the selection of a solvent system with a givenvalueof E; that is,once the curve has been established, the optimum water /solvent ratio for another solvent can be calculated from known dielectric constant relationships (Cavi: et al., 19'79).

Frequently, ternary solvent systemsare resorted to. Examples are water-propylene glycol-benzyl alcohol or water-propyleneglycol-ethanol. In suchcases the solubility profile is usually presentable by a ternary diagram (Sorby et al., 1963). This type of diagram usually demands a fair amount of work; that is, the solubility of the drug substance in many solvent compositions must be determined. A priori, it would therefore seem that they would be out of place in a situation where only limited quantities of drug are available. However, their principle gives some validity to optimization procedures. The diagram can be of one of two types, as shown in Figs.4 and 5. In the first type, the solubility may be assumed to be of the type S = a10

+ all + a12C2 C1

(9.5)

where C denotes concentrations of nonaqueous solvents. An example of this is ere the subscripts to C denote the two nonaqueous solvents. Hence three solubility experiments woulddetermine the relationship (with zero degrees of freedom). It is usual to do at least five, and determine possible curvature [i.e., inclusion of more terms in Eq. (9.6)]. a parabolic type curveas shown. In the second case,Fig. 5 each tie line will give Hence at a given concentration of C2 the solubility can be approximated by

Ternary diagram of solubility of a compound in a ternary mixture with linear constant c on cent ration ity response,Inset: Concentration of drug in compositions with of B. The composition of the solute is theconstant concentrationof ,the concentration of A in the abscissa, and the complement concentration of C (the third apex, not indicated in the figure). The drug solubility response is linear in the A concentration in this case.

\ \

Y O

Ternary diagram and tie line concentration in a nonlinear system.

where, in the simplest case, h 0

= c20

+ c21c2

(9.7)

timization can be achieved by five (or more experiments, with zero (or in general n-5) degrees of freedom.

It is advanta~eous,with a new drug substance, to be able to estimate what its solubility might be, prior to carrying out dissolution experiments. There are several systemsofsolubility prediction, e.g., the work by Amidon et al. (197 and ~alkowsky et al. (1972, 1975). Their equation, for solubility ndmixedsolvents,is a simplifiedtwo-dimensional debrand equation and isbased on the product of molecular surface area of the hydrocarbon portion of a molecule. odor et al. (1989) have developed a semiempirical solubility ariables (S=molecular surface in. A2, 1, =indicator variable Qn=square root of sum lculated dipolemomentebyes, on oxygen atoms, root of sum of squared charges =molecular volu =square of molecular surface, =sumof =molecular weight, (0)= ovality of molecule, absolute values of atomic charges on hydrogen atoms, Ab, = sum of absolute values of atomic charges on carbon atoms, A,, =indicator variable for aliphatic amines, single bonds in the molecule). and i v h =number of The aqueous solubilities, W9 of331 compounds were found to follow the equation (with tolerances omitted)

+ 0.32235~-0.591431, + 38.443 + 34.569Q: -31.8 + 1.9882Am+ 0.15689ivh + 0.00014102S2+ 0.40308s -0.59335Ab~ 2352V + 1,3168Abh + 108.80{0}-61.272{ 39

(9.8)

f the parameters listed onlythe ovality and the indicator unfamiliar entities that are obtained from the literature

ue for the alkanes I, are odor et al., 1989).

The importance of dissolution is such(from biopharmaceutical considerations) that used now is throughout As ondosage solid forms. ccording to Noyes and *

dm dt

VdC --kA(S -6> dt

" " __.

where m is massnot dissolved, Vis liquid volume, t is time,k is the so-called intrinsic dissolution rate constant (cmls), and A is surface area of the dissolving solid.

50 l

40 30

10 0

100

Time t Dissolution profiles obtained fromthe solubility determination of two polymorphic forms of the same drug substance. A is the stable form with solubility 31 mg/mL. B is the profile of the metastable form with solubility 46mg/mL. This solubility, C, (circles) is not achieved in many instances, and precipitation of the stable form occurs at a point beyond the solubility of A, and the trace becomes B.

Many criticisms have been voiced against Eq. (9.9), but in generalit is correct, and it will be assumed to be so in the following. Experimentation can be carried out with constant surface as when using a Wood’s apparatus (Wood et al., 1963) or, with smaller amounts, making a small pellet and encasing it in wax and exposing only one face to a dissolution medium, or simply employing an excessof solid throughout the dissolution experiment. In such cases Eq. (9.9) may be integrated to give (9.10) or .=,Cl

-exp(-~t)]

(9.11)

A typical curve following Eq. (9.11) is shown in Fig. 6. In the critical time path for product development, solid dosageforms (tablets or capsules) must eventually be manufactured for the clinic (e.g., in clinical phase 11). Ifpossible, the drug substance per se is subjected to a dissolution test in a Wood’s apparatus (Wood et al., 1963). This test is useful although it is quite dependent on hydrodynamic conditions. It consists of placing the powder in a special type of tablet die, compressing the tablet, and exposing the flat, exposed side of the tablet (with surface area A) to a dissolution liquid (usually water or N/10 has a solubility S. Under these conditions (Carstensen, 1974), the intrinsic dissol-

ution rate constant (cm/s) can be obtained by Eq. (9.10), which under sink conditions (i.e., when C is less than 15% of S) becomes

C=- SkA V

(9.12)

It has been suggested(~iegelman,1979) that if k is obtained under sink conditions over a pH range of 1 to 8 at 37°C in a USP vessel by way of Eq. (9.12) at 50 rpm9 then if the dissolution rate constant (kA/ v> is greater than 1mgmin" cm-29 the drug is not prone to give dissolution-rate-limited absorption problems. On the other hand, if the valueisless than 0.1,suchproblems can definitelybe anticipated, and compounds with values of kA/ V of from 0.1 to 1mgmin" cm-2 are in a gray area. For compound selectivity it is frequently usefulto express dissolution findings in terms of k (i.e., in cm/s). For a small amount of powder, dissolution of the particulate material can often be assessed(and compared with that of other compounds) by placing the powder in a calorimeter (Iba et al., 1991) and measuring the heat evolved as a function of time. The surface area must beassessedmicroscopically (or by image analyzer), and the data must be plotted by a cube root equation (Hixson and Crowell, 1931): 2kS 1 - [ ~ ]'l3=-pr

(9.13)

where.M is massnot dissolved, MO the initial amount subjected to dissolution, p true density, S solubility, and r the mean c c r a d i ~of~ the 7 7 particle. The method is simply comparative, not absolute, because the hydrodynamics are different in the calorimeter from what it would be in a dissolution apparatus. It is obvious that the dissolution rate is a function of the exposed surface area, but how this changes during dissolution is not quite obvious. Sunada et al. (1989)measured the change in surface area during dissolution of n-propylp-hydroxybenzoate and found dissolution rates proportional to surface area.

Quite often a compound is rather unstable in aqueous solution. Hence the long exposure to liquidrequired for traditional solubilitymeasurementswillcause decomposition, and the resulting solubility results will be unreliable. In this particular case Nogami's method may be used. Ifa solution experiment iscarried out as a dissolution experiment with samples taken at equal time intervals, 6, it can be shown (Nogami et al., 1966) that when the amount dissolved at time t 6 is plotted versus the amount dissolved at time t, a straight line willensure. The following relationship holds:

C(t

+ S) = SI1 -exp(-kS)] + exp(-kS)Ct (9.14)

hence such a plot as shown in Fig. 7 will give k from the slope; inserting this in the S. The advantage of the method is that itcan be carried intercept expression will give out in a short period of time, and reduce the effect of decomposition; the disadvantage is that it is not as precise as ordinary solubility determinations.

c at t

The Nogami method appliedto the upper curve Fig. in 5. The l~ast-squaresfit for the upper line is y=27.081+0.53139~, so that y = x when x=y=22.081/(1-0.53139)=47.3. ili

lymorphism isan important aspect of the physical properties of drug substances. ne of the characteristics of a metastable polymorph is that it is more soluble than its stable counterpart. The dissolution profile of it will b as the upper curve shown in Fig. 6; A is the stable form with solubility 31 mg/mL. is the profile of the metastable form with solubility*46 mg/mL. This solubility (c les) is often not achieved, and precipitation of the stable form occurs at a point beyond the §olubility of A, and the trace becomes B. In such. cases, the Nogami method can be appliedto the early points curve (Fig. 7), and the solubility, S‘, of the polymorph can be assessed. One of the important aspects of metastable polymorphs in pharmacy isexact their higher solubility, since the dissolution rate will also be higher [Eq. (9.9)]. ence the bioavailability will be increased where this is dissolution rate limited (Shibata et al., 1983).

As mentioned at anearlier point, solids may exist either as amorphous compounds or

as crystalline compounds. In the latter, the molecules are positioned inlattice sites. A lattice is a three-d~mensionalarray, and there are eight systemsknown. ~ o m ~ o u n d s often have the capability of existing in more than one crystal form, and this phenomenon is referred to as polymorphism. If a compound exhibits polymor~hi§m,one of the forms will be more stable (physically) than the other forms; that is, ofn existing forms, n-l forms will possess a thermodynamic tendency to convert to the nth, stable form ( ~ h i c hthen has the lowest Gibbs energy; it should be noted that in the preformulations stage it is not known whether the form on hand is the stable polymorph or not). * It is noted that the “solubiLity” is not the e q ~ l i b r solubility. i~ The solution is a supersaturated solution, but it is referred to as solubility, because conducting a solubility experiment on a metastable most often will give a reproducible figure. The s u p e r ~ a t ~ a t esolution d will eventually precipitate out as demonstrated in Fig. 5 of Chapter 12.

e manner in which different polymorphs are created is bywayof lizing them from different solvents, and at a point in time when sufficient quantities of material (and this need not beverymuch) are available, the ation scientist should undertake recrystallization from a series of solvents. wledge of polymorphic forms is of importance in preformulation because suspension systemsshould never be made with a metastable form (i.e., a form other than the stable crystal form). Conversely, a metastable form is more stable than a stable modification, and this can be of advantage in dissolution [Eq. (9.1 l)].

In general, vapor pressures are not all that important in preformulation, but it should always be kept in mind that a substance may have sufficiently low vapor pressure to (a) become a lost to sufficient extent to cause apparent stability problems and content uniformity problems, and (b) exhibit a potential for interaction with other compounds and adsorption onto or sorption into package components ( and Lukes, 1976). ost drug substances are, substantially, not volatile. As an initial screen, it can mined whetherthe drug is sufficientlyvolatile to cause concern, by placing a weighed amount of it in a vacuum desiccator and weighing it daily for a while, It is better to have a high-vacuum systemfor this, and the use ofa vacuum electrobalance is best for this p pose. A good estimate of the vapor pressure can be obtained (Carstensen and thari, 1981)byusing a pierced thermal analysis cell,placing it on a vacuum electrobalance, and monitoring the weight loss rate. A substance with known vapor pressure can then be used for calibration, the loss rates being proportional to the vapor pressures. y using constant temperature TGA, graphs such as that shown in Fig. 8 will result. The weight rate (which should be established as due to evaporation of the compound) is given by d dt

--kA'P,

(9.15)

1 .o

0.9

0.8

Q" 0.7

Q" 0.6

eight loss curves from constant temperature TGA.

2.7

2.8

2.9

3.0

3.1

l ooofr Plot showing the Clausius-Clayperon treatment of the data in Fig. 7. The least-squares fit is In[-slope] =22.861-8.7(1000/T.)

where Wis weight, t is time,k is a heatlmass transfer coefficient, A' is surface area, and P a is the vapor pressure of the compound at the given temperature. If the specific surface area is

where Wao is the original weight of the sample, we inserted this into Eq. (9.5) and obtain (9.17) An example of such a plot is shown in Fig. 8. If a compound with known vapor pressure (e.g., benzoic acid, subscript b) is subjected to the same conditions, then it will exhibit a weight loss curve given by a similar equation: (9.18) The ratio of the slopes isA b p b / k f a P a , so if the specific surface area is knownfor each, then (since P b of the reference is known), P a can be calculated. The heat of sublimation, AH, can be obtained by plotting the negative of the slopes by a Clausius-Clayperon equation, as shown in Fig. 9. In the cited case, the negative of the slope of this plot is A H l R = 8.7 kCallmo1, so that the heat of sublimation is A H = 17.4 kCallmo1. For this, no reference is necessary;it is only necessary to know and adjust for the weights of the samples studied.

Partition coefficients between water and an alkanol (e.g., octanol) should be determined in preformulation programs (Yalkowsky et al., 1983). The partition coefficient of a compound that exists as a monomer in two solvents is given by (9.19) If its exists as an n-mer in one of the phases, the equation becomes (9.20) or

1 -log c2 logk = nlog c

(9.21)

The easiest way to determine the partition coefficient is to extract V1 cm3 of saturated aqueous solution with V2 cm3 of solvent and determine the concentration C2 in the latter. The amount left in the aqueous phase is C1V1 -C2V2 = the partition coefficient is given by (9.22) If it is assumedthat the species is monomericin both phases, the partition coefficient becomes the ratio of the solubilities, and it is simply sufficient to determine the solubility of the drug substance in the solvent (sinceit is assumedthat the solubility is already known in water): (9.23)

Hygroscopicity is, of course, an important characteristic of a powder. It can be shown for a fairly soluble compound that the hygroscopicityis related to its solubility (Carstensen, 19’7’7, VanCampen et al., 1980), although it has been shown that the heat of solution plays an important part in what isconceived as “hygroscopicity”(VanCampen et al., 1983a,b,c). As mentioned in Chapter 8, a hygroscopicity experiment is carried out most easily by exposing the drug substance to an atmosphere of a known relative humidity (e.g., storing it over saturated salt solutions in desiccators). Each solution will give a certain relative humidity (RH), and the test simply consists of weighing the powder from time to time and determining the amount of moisture adsorbed (weight gained). This does not work with drug substances that decompose as, for instance, effervescent mixtures, which start losing weight due to carbon dioxide evolution (Carstensen and Usui, 1984).

100

Relative H u ~ i d i t y Socption isothermof anhydrous droloxifene citrate. (Graph constructed from data pub~ishedby Burger and Lettenbichler, 1993.)

It can be shown that if the air space is sufficiently agitated to prevent vapor pressure gradients, the initial uptake rate (g H2O/g solid per hour) is related to the relative humidity by

H0 is the vapor pressure of a saturated solution of the drug substance in water. An example of this is shown in Fig. 10. X, can be estimated by an ideality assumption; that is, if the solubility is expressed as a mole fraction .Xs, the vapor pressure over a saturated solution will be P’ given by

P’= (1 -X,)PO

(9.25)

where PO is water’s vapor pressure at that temperature. The experiments above are rather easy to carry out and should always bepart of a preformulation program, since hygroscopicitycan be so important that it will dictate whether a particular salt should be used. Dalmane, for instance, is a monosulfate, and is usedas such sincethe disulfate, desirable in many other respects, is so hygroscopic that it will remove water from a hard-shell capsule and make it exceedingly brittle.

Prior to attempting the first formulation with a new drug, most researchgroups carry out compatibility testing (Carstensen et al., 1964). The principle is to make up reasonably ratioed mixtures of drug and excipient, to ascertain which excipients maybe reasonably usedwith the drug. original method used in the 1960s mixtures, (Carstensen, 1964) consisted visof observation of such spectrophoto~etricassay, and TLC. The methods used nowadays have followed with analytical developments and are (a) chemical assay, (b)TLC, (c) C, and (e) microcalorimetric methods. The latter twohavebeen of 1 interest in recent years and will be treated separately.

ref

Rustichelli et al. (1999) have employed D S c to obtain the phase equilibrium diagrams of the enantiomers of (a) verapamil Cl and (b) gallopamilHCl. In the former case the eutectic composition is at 90% (2s)-(-)-~erapamilHCl and in the latter at 70% (2S)-(-)-gallopamil HCl. ra et a1 (1998) have usedthermal analysis (DSc) tostudy compatibility of n pharmaceutical excipients (palmitic acid, stearic acid, ,000, and sorbitol) and showed that the interactions were primarily due to dissolution in the melted excipient.

Heat conduction microcalorimetry has been used as a method to evaluate stability and excipient stability by a series of researchers. Angerg et al. (1988, 1990, 1993), ansen et al., (1989), and Wilson et al. (1995) have described the general method and results interpretation, For instance, Angberg et al. studied the oxidation of ascorbic acid in aqueous solution by microacalorimetry, and other researchers have used this method as well. Oliyai and Lindenbaum (199 1)studied the decomposition of ampicillin in solution. Tan and Meltzer (1992) studied the solid state stability of 13-cis-retinoic acid by means of microcalorime6ry and HPLC, and Pika1 and Dellerman (1989) studied the kinetics of cephalospop~in inthe solid'& states using the same method. Seltzer et al. (1998) used the method to evaluate stability and excipient compatibility of

(~)-(3-(2-4-(S)-(4-(amino-imino-methyl)-phenyl-4-methyl-2,5-dioxo-imidazolidin-l-yl)-acetylamino))-3-phenyl-propionicacid ethyl ester, acetate. The excipients used were potato starch, calcium hydrog phate anhydrous, and colloidalsilica. D and denote the concentration of C as They consider the reaction A x,the fraction decomposed at time t. The trace of heat evolved as a function of time is then characterized by

dx = --k{[Ao] -X I n -

(9.26)

The amount decomposed isassociated with (a usually exothermic)reaction enthalpy, AH, where the heat evolved isproportional to x,hence the heat flow isproportional to dxldt and the proportionality constant is AH, so that dt

cf,=CIQ=AHdt

(9.27)

hence (9.28) For a first-order reaction this becomes (9.29)

-10

-20 0.0

1 .o Time(days) Figure constructed from data published by Seltzer et al. (1998).

so that the terms k and AH may be deduced from nonlinear regression of a plot of heat flow versus time (Fig. 1l), If the reaction is zero order, then # is simply a constant:

# =k A ~

(9.30)

This allows determination of kinetic profiles andArrhenius plots of the studied reaction. The value of Q at time t is obtained through integration (area under) of the curve from zero to t . It should again be emphasized that at the onset of a new drug program, there are only small amounts of drug substance at hand. One of the first tasks for the preformulation scientist is to establish the framework within which the first clinical batches can be formulated. To this end it is important toknow with which common excipients the drug is compatible. In the following, the distinction willbe made between solid and liquid dosage forms. The microcalorimetric methods give no direct information about the chemical nature of the reaction. OS

It is customaryto make a small mix of drug substance with an excipient, place it in a vial, place a rubber stopper in the vial, and dip the stopper in molten carnauba wax (to render it hermetically sealed). The wax will harden and forma moisture barrier up to 70°C. A list of common excipients characteristic of this type of test is shown in Table 1. At times it is possible to obtain quantitative re1 ionships of excipient characteristics and interaction rates (Carstensen et al., 1964; rrier and Kesselring, 1983). In addition to the test as described, a similar set of samples are set up where 5% moisture is added. A storage period of 2 weeks at 55°C is employed [except

egories for Two-Component Systems Worse 17-27

Identmo 25°C ical Drug per se

+ Magnesium stearate + Calcium stearate + Stearic acid + Talc + Acid-washed talc +Lactose + CaHP04, anhydrous + Cornstarch +Mannitol + Terra alba + Sugar 4 x

15 9 16 15 13 12 15 7 14 10 12 10 12 9 12 9 12 10 10 8 14 11 12 9

at 4 8 3 4 4 5 5 11 5 8 8 9 5 7 6 8 5 5 7 7 6 6 6 7

Total score

10 days at 55°C

1 3

1 1

3 3 0 2 1 2

0 1

3 4 2 3 3 5 3 5 0 3 2 4

25"

55°C

38 49 34 43 37 38 42 60 38 45 44 49 38 65 46 66 39 40 39 47 41 50 41 63

31 38 30 35 32 35 31 38 30 34 31 35 32 56 36 53 34 37 31 45 28 45 34 61

Source: Constructed from data published by Carstensen et al. (1964).

for stearic acid, where 45°C is used, and dicalcium phosphate, where 37°C is used (Toy, 1980)], after which time the sample is observed physically for (1) caking, (2) liquefaction, (3) discoloration, and (4) odor or gas formation. It is then assayed by thin-layer chromatography (or HPLC). It is noted that one of the samples setup is the drug by itself. This is done for several reasons, one of whichis that it is now required by the FDA for IN submissions (Schultz, 1984). One more reason is that at the onset of a program, the organic synthesis of the compound may lack the refinement it will later have, and it is not uncommon that there will be several weak TLC spots (impurities) on a TLC chromatogram of a compound obtained by initial laboratory synthesis. Hence, in selecting the excipients with which the drug substance is deemed to be compatible, it is customary to use as criteria that (after accelerated exposure of a druglexcipient mix) no new spots have developed and that the intensity of the spots in the drug stored under similar conditions (2 weeks at 55°Cis the same as in the acceptable excipient. This type of program is used by many companies

with good success (i.e.,the formula developed based on the findings from the compatibility program is stable). It should be noted that liquefaction at times occurs because of eutectic formation (e.g., often with caffeine combinations) and that this may not necessarily be associated with decomposition. On the other hand, discoloration (e.g., amines and sugars) usually is. Finally, the reason for not forcing dicalcium phosphate (a very valuable formulation aid in direct compression) beyond 50°C is that at higher temperatures it converts to the anhydrate, a qonversion that is,curiously enough, catalyzed by water. In other words, the dihydrate willbe autocatalytic in this respect at elevated temperatures, and it should not be ruled out based on high-temperature findings. Aside from magnesium stearate, dicalcium phosphate and lactose are the excipients that are the most often found incompatible with drugs. In the former case it is usuallythe pH effect, in the latter itis the formation of Schiff‘s bases with amines (and many drugs are amines), i.e.,

For instance, Eyjolfsson (1998) reported on the incompatibility of lisinopril with lactose. Aso et al. (1997) have determined the decomposition rates of cephalotin in mixtures with pharmaceutical excipients and the effect of moisture. They found a linear relation between mobile water percentage and decomposition rate constants. Several examples of this type of screening exist. Malan et al. (1997) have studied the compatibility of tablet excipients with albendazole and closantel. They prepared drug-excipient mixtures in a mixture and in 1 :1 mi~turesthat were granulated with water and dried at 50°C. DSc and HPLC were used to evaluate tested were colloidal silicon dioxide, the compatibilities. The . excipients microcrystalline cellulose, dibasic calciumphosphate monohydrate, starch, sodium starch glycolate, and magnesium stearate.

Compatibility studiesmay also include compatibility with container materials. Hourcade et al. (1997), for instance, reported that granisetron in concentrations of 1mg/mL, when kept in polypropylenesyringes,were quite stable, whereas dilutions with 0.9% NaC1 or with 5% glucose resulted in unsatisfactory storage stability. FIL

les have been discussed in Chapter 3. Frequently, a broad screen of stability med on the initial small sample used for initial preformulation; this is frequently referred to as “forced decomposed studies” (Bodnar et al., 1983). In this the drug isexposed to “acid degradation,” “base degradation degradation,’’ “drug powder degradation,” and “light degradation.” studies are eventually needed.

For any compound marketed by a pharmaceutical concern, at one time during its development, there should be a concerted project to establish a very Q this correctly is a time-consumin the information that can be gleaned from it isvery important with regard to ulations, and it is therefore customary to carry out an approximate kinetic rofile carst tense^ et al., 1992) early in the development stage. This will allow formulation of solutions for injections and fo using buffers, that willgive the best stability. thout it formulation is essentially guesswork.

ortant part of liquid compatibilities. ver, and p in aqueous (or other types of)sol 2. This is nowrequired in the stability guidelines, which state that “it is suggested that the following conditions ... be evaluated in studies on solutions or suspensions of bulk drug substances: acidic ,high oxygen and nitrogen atmospheres, and the presence of adde such as chelating agents and stabilizers” and it is suggested“that stress testing conditions ... include variable temperature (e.g., 5, 50, 75’C).’’

In general, such studies are carried out by placing the drug in a solution of the additive. These can be (and usually are) a heavy metal (with or without chelating agents present) or an antioxidant (in either oxygen or nitrogen atmosphere). Usually, both flint and amber vials are used, and in many cases an autoclaved condition is included. This willanswer questions about susceptibility to oxidation, to light exposure, and to heavy metals. These are important questions as far as injectabl compatibilities are concerned.Exposure to various plugsis frequently included at this point so that early injectable preparations can be formulated. For preparations for oral use,knowledgeof the desireddosage form is important, but compatibility studies withethanol, glycerin, sucrose,corn syrup, preservatives, and buffers are usually carried out. This type of study also gives an idea of s for the activation energy, ,of the predominant reaction in solution. ~ r r h e n i uplots compounds in solution are usually quite precise.

With transdermal dosage forms being of great importance of late, it is advisable to test for compatibilities with ointment" excipients and withpolymers(e.g., ethylvinyl polymer, if that is the desired barrier). In the case of transdermals, the dosage form is either directly placed in a stirred liquid or it is placed in a cell with an appropriate membrane (e.g., Cadaver skin) to estimate the release characteristics of the drug from the ointment (Chien et al., 1983). It should be noted here that if the overall flux is J,then 1

- 1

“

Jointment

1 Jmembrane

(9.3 1)

Time (Arbitrary Units) Slope determination of ointment +-membrane.

flux of

ointment release

and

release from

where subscripts refer to the respective phaseJmembraae can be obtained from curves such as shown in Fig. 12 in the fashion that first the overall flux is obtained (with the membrane in place), giving the value of J, then the release is obtained without with membrane in place, giving Jointment, that is

l dm1 A dt

J-” __.

(9.32)

and Jointment

I==

1 dm2 Adt

(9.33)

is then obtained as the reciprocal of the difference. In vivo testing is usually carried out by applying the dosage form to hairless rats followed by subsequent sacrifice. Since the skin consists of a number of layers with differing hydrophilicity, the overall fate of the drug is of importance. Jmembrane

In the case of emulsions, the preformulation studies becomevery formulation oriented. Williams and Mahaguna (1998) have described preformulation studies ofFreund’sincomplete adjuvant (FIA), whichis a water-in-oilemulsion. This included measuring the critical micelle concentration of the formulations to be investigated. Using ovalbumin (a model antigenic protein) in the interface, the surface activity of mannide monooleate, in the interface between water and oil phases, was determined.

Wong etal. (199’7) studied the stability of cefazolin in Pluronic F-12’7 gelsand found the decomposition to be first order, and for all Pluronic concentrations used (20,25, and 30%), Arrhenius plotting was satisfactorily linear.

Acree, W. E., Rytting, J. H. (1983). J. Pharm. Sci. 72:293. Amidon, G. L., Yalkowsky, S. H., Leung, H. (1974). J. Pharm. Sci. 63:1858. Angerg M., Nystrom, C., Castensson, S. (1988). Acta Pharm. Suec. 25:307. Angerg M., Nystrom, C., Castensson, S. (1990). Int. J. Pharm. 61:66. Angerg M,, Nystrom, C., Castensson, S. (1993). Int. J. Pharm. 90:19. Aso, Y., Sufang, T., Yoshka, S., Kojima, S. (1997). Drug Stability 1:237. Bakar S. K., Niazi, S. (1983), J. Pharm. Sci. 72:1024. Bodnar, J. E., Chen, J. R., Johns, W. H., Mariani, E. P,, and Shinal, E. C. (1983). J. Pharm. Sci. 72:535. Bodor, N., Cabanyi, Z.,Wong, C.-K. (1989). J. Am. Chem. Soc. 111:3783. Burger, A., Lettenbichler, A. (1993). Eur. J. Pharm. Biopharm. 39:65. Carstensen, J. T. (1974), inDissolution Technology (Leeson,L., Carstensen, J. T., eds,), Academy of Pharmaceutical Sciences. American Pharmaceutical Association, Washington, DC, p. 5. Carstensen, J. T. (1977). Pharmaceutics of Solids and SolidDosage Forms, New York: Wiley-Interscience, pp. 11-1 5. Carstensen, J. T., Kothari, R. (1981). J. Pharm. Sci. 70:1095. Carstensen, J. T., Usui, F. (1984). J. Pharm. Sci. 74:1293. Carstensen, J. T., Johnson, J. B., Valentine, W., Vance, J. (1964). J. Pharm. Sci. 53:1050. Carstensen, J. T., Su, K. S., Maddrell, P., Newmark, H. (1971). Bull. Parenter. Drug Assoc. 25:193. Carstensen, J. T., Franchini, M., Ertel, K. (1992). J. Pharm. Sci. 81:303. Cavk. G., Puisieux, F., Carstensen, J. T,, (1979). J.- Pharm. Sci. 68:424. Chien, l’. W., Keshary, P. R., Huang, Y. C., Sarpotdar, P. P. (1983). Drug. Dev. Ind. Pharm. 72:968. Chollet, J. L., Jozwiakowski, M. J., Phares, K. R., Reiter, M. J., Roddy, P. J., Schultz, H. J., Ta, Q. V., Tomail, M. A. (1999). Pharm. Dev. Tech. 4:35. Eyjolfsson, R. (1998). Drug Dev. Ind. Pharm. 24:797. Hansen, L. D., Lewis, E. A., Eatough, D. J., Bergstrom, R. G., DeGraft-Johnson, D. (1989). Pharm. Res. 6:20. Hixson, A., Crowell, J. (1931). Ind. Eng. Chem. 23:923. Hourcade, F., Sautou-Miranda, V., Normand, B., Laugier, M., Picq,F., Chopineau, J. (1997). Int. J. Pharm. 154:95. Iba, K., Arakawa, E., Morris, T., Carstensen, J. T, (1991). Drug Dev. Ind. Pharm. 17:77. Malan, C. E. P, devilliers, M. M,, Lotter, A. P. (1997). Drug Dev. Ind. Pharm. 23:533. Mura, P., Faucci, M. T., Manderioli, A., Furlanetto, S., Pinzauti, S. (1998). Drug. Dev. Ind. Pharm. 24:747. Nogami, H., Nagai, T. Suzuki, A. (1966). Chem. Pharm. Bull. 14:329. Noyes A., Whitney, W. (1897). J, Am. Chem. Soc. 23:689. Oliyai, R., Lindenbaum, S. (1991). Int. J. Pharm. 73:33. Parke, T.,Davis, W, (1954). Anal. Chem. 25:642. Paruta, A. N., Irani, S. A. (1964). J. Pharm. Sci. 54: 1334. Perrier, P. R., Kesselring., U. W. (1983). J. Pharm. Sci. 72:1072.

rst

Pikal, M. J., Dellerman, K. M. (1989). Int. J. Pharm. 50:233. Pikal, M., Lukes, A. L. (1976). J. Pharm. Sci. 65:1269. Riegelman, S. (1979). Dissolution Testing in Drug Development and uality Control, The Academy of Pharmaceutical Sciences, Task Force C o ~ i t t e e American , ~harmaceutical Association, p. 31. odriguez-Hornedo, N. (1984). Crystallization Kinetics and Particle Size thesis, University of Wisconsin. Rustichelli, C., Gamberini, M. C., Ferioli, V., Gamberini, G. (1999). Int. J. Pharm. 178:lll. Schultz, R. C. (1984). Stability of Dosage Forms, FDA-Industry Interface Washington, D. C., Oct. 7, 1983. Selzer, T., Radau, M., Kreuter, J. (1998). Int. J. Pharm., 171:227. Shibata, M., Kokobu, H., Morimoto, K., Morisaka, K., Ishida, T., Lnoue, Pharm. Sci.72:1436. Sorby, D., Bitter, R,, Webb, J. (1963). J. Pharm. Sci. 52:1149. ,Congressional Record, May 7, 1984. D. H.-S. (1998). Int. J.Pharm. 164: 139. ara, I., Qtsuka, A., Yonezawa, Y. (1989). Chem Pharm. Bull. 37:467. Tan, X., Meltzer, N. S. L. (1992). Pharm. Res. 9: 1203. Toy, A. D. F., (1980), Inorganic phosphorous chemistry. In Comprehensive Inorganic Chemry (J. C. Bailar, Jr., H. J,Emelius, R. Nyholm, A. F. Trotman-Dickenson, eds.), A. heaton, Exeter, UK, pp. 389-543. rg, W. J. W., Lingeman, H. (1983). J. Pharm. Sci. 72:553. Van Campen, L., Zografi, G., Carstensen, J. T. (1980). Int. J. Pharm. 5:l. Van Campen, L., Amidon, G. L., Zografi, G. (1983a). J. Pharm. Sci. 72: 138 1. Van Campen, L., Amidon, G. L., Zografi, G. (1983b). J. Pharm. Sci. 72:1388. Van Campen, L., Amidon, G. L., Zografi, G. (1983~).J. Pharm. Sci. 72: 1394. .Q, 111, Mahaguna, V. (1998). Drug Dev. Ind. Pharm. 24: 157. seman, G. (1954). J. Physiol. 123:116. zer, A. E., Mitchell, J. C., Loh, W. (1995). J. Phys, Chem. 99:7108. Wong, C.-Y., Wang, D.-P., Chang, L.-C. (1997). Drug Dev. Ind. Pharm. 23:603. Wood, J. H., Catacalos, G., Lieberman, S. (1963). J. Pharm. Sci. 52:296. Yalkowsky, S. H., Valvani, S. C. (1983). J. Pharm. Sci, 72:912. Yalkowsky, S. H., Flynn, G. L. and Amidon, G. L. (1972). J. Pharm. Sci. 61 :983. Yalkowsky, S. H., Amidon, G. L., Zografi, G., Flynn, G. L. (1975). J. Pharm. Sci. 64:48. Yalkowsky, S. H., Valvani, S. C., Roseman, T. J. (1983). J. Pharm. Sci. 72:866. *

adi is on, ~ i s ~ o n ~ i n

1. Physical Stability of Solutions 1.1. Organoleptic testing 1.2. Subjective appearance testing

263 263 264

2. Parenteral Solutions 2.1. Swirly precipitates 2.2. Whiskers 2.3. Cloud times 2.4. Oral solutions

264 265 266 266 268

3. Disperse Systems 3.1. Suspensions 3.2. Sedimentation volumes 3.3. Sedimentation rates 3.4. Preservation stability 3.5. Dissolution of suspensions 3.6. Temperature testing of dispersesystems 3.7.Semisolidsuspensionsystems (ointments, suppositories) 3.S. Ointments and transdermals

268 269 27 1 273 273 274 274 274 275

4. Emulsions 4.1. The emulsion interface 4.2. Globule size and viscosity 4.3. Stability of the emulsifier/protective colloid system 4.4. ~mulsiontype 4.5.Rheological properties 4.6. Appearance of emulsionsystems 4.7. Breaking and coalescence 4.8.Semisoliddosage forms

276 276 277 280 280 28 1 282 282 283 1

4.9. Transdermals

5. Accelerated Testing and Prediction

283 285

Aerosols 6. 6.1.Aerosol testing 6.2. Sprays

286 287 289

7. Powders

289

293 293 296 299 300 303 306 306 307 307 308 312

Tablets 8.1. Tablet hardness Softening 8.2. 8.3. Disintegration 8.4. Porosity of tablets 8.5. Dissolution 8.6. Percolation thresholds 8.7. Multipoint determinations 8.8. Dissolution media 8.9.In-vivo to in-vitro correlation 8.10. Stability of dissolution curves 8.11. Appearance of tablets and capsules

9. Sustained Release Products 9.1. Coated beadlets and granules 9.2. Erosion tablets 9.3. Insoluble matrices 9.4. Osmotic pump 9.5. Gel forms

314 314 316 316 318 318

10. Coated Tablets 10.1. Film coated tablets 10.2.Sugar coated tablets 10.3. Enteric coated tablets

319 319 320 320

11. Hard and Soft ShellCapsules

321 322

13.LightSensitivity

Testing

14. Diagnostic Papers

322 324 324

References

325

A great deal of space has been devotedto the subject of chemical testing. However, even if a product, chemically, is sufficiently stable to sustain e.g. a 3-year expiration date, physical changes may have occurred. In a solid dosage form, the dissolution may have slowed down to such an extent that the product is no longer as bioavailable

as it was at the time of manufacture, and more importantly, it may not meet the minimum required for efficacy. For a solution, a precipitate may have occurred. This may not affect the chemical content, but for a parenteral product it would, obviously, be quite unacceptable, and for an oral solution it would also be unsatisfactory, because the dispensing pharmacist would rightfully question the integrity of the product. The caking of a suspension impairs the dispensing of a known amount of drug in a teaspoon, and a separated or broken emulsion or cream obviously will not have the same emollient properties as would a proper product. Physical stability will betreated by product category in the same order as in the case of chemical stability.

Solutions are broadly divided into two categories: oral and parenteral solutions. Appearance, in both cases, is an important factor. In the case of oral solutions, organoleptic properties are also of great importance. Organoleptic evaluation is usually done subjectively, i.e.,a tester (operator, technician), will judge the product and score it, either numerically or descriptively or both. In the case of appearance of solutions, there should always bea subjective statement (quantitative or subjective description) even ifmore quantitative instrumental parameters are recorded. A few words are therefore in order regarding organoleptic and appearance testing.

For organoleptic testing it is important to establish a test panelearly in the stability program. (Or if a stability program is in place, but no such testing is carried out, a test panel should be selectedat the first opportunity when a product with important taste or odor properties is placedon stability.) Many companies utilizejust one tester for the task of organoleptic testing, but this can be shortsighted, because the tester may leave, go on vacation, or become ill, and in that case the logical solution is to assign someone else to the task. There may be an evaluational bias between the two testers, and this should be established at the onset. First of all, the depth of organoleptic capacity should be tested. This can be by asking the tester to taste serial dilutions of a bitter substance (e.g., quinine). e a sensitivity levelcan be established. A control of e.g. water or high dilutions should always be part of the protocol. It should be noted that the technicians are not taste testers in the ordinary sense. That is, it is not necessary to match their “likings” to that of the general public. Rather, itis important thatthey can (a) duplicate their results and (b) remember them, since they willbe asked to taste a preparation that they originally tested3 or 6 months earlier. In so doing they would have to score the degree of flavoring, e.g., is it less than originally present, i.e., is the flavor being lost? They would also have to be able to describe the flavor well originally. For example, if the chemical is slightlyanesthetizing, the duration of the anesthesia would be important. If there is interaction with a plastic bottle, are off flavors appearing in the product? Finally it is important toscreen severaltesters to ascertain that they givethe “same result.” In describing the flavor, several categories can be used (degree of sourness, degree of saltiness, level of flavor, type of flavor). Each of these may be assigned

to a level of e.g. 1-5. A flavor profile may hence be established, and this can then be reestablished at several time points in the room-temperature storage. It is not recommended to evaluate results from higher temperatures (although they maybe carried out).

tions, particularly parenteral solutions, may havea tendency to discolor slightly. n it is not possible, withinanalytical sensitivity, to establish either the source of e color or the level of the substance causing it, In this case it is a good practice to use a color standard to describe the “intensity” of the discoloration. for instance, uses the so-called Roche Color ~tandard(RCS), which uses pound (the identity of which is a secret) that can be reliably reproduced and has exceptional color stability. Making up serial dilutions of this CO solutions of different “slight” discolorations; they are denoted so that a solution can alwaysbe compared in this fashion, old-fashioned Dubosque colorimeter (which can be used withadvantage in this type of situation). The prin~ipleof the ~ u b o s ~ colorimeter ue is to have a view of two test tubes from the top. One isthe control, and the other is the solution being matched. It is possibleto adjust the length of the light path in the second tube, and this is done until the intensity matches that of the standard. The length of the path is then an indication of the “concentration.” CS (and similartypes of numbers) are difficult to analyze, but a Dubosque colorimeter gives numbers that follow Beer’s law and are logarithmically proportional to concentration (although the proportionality factor cannot be known). In this fashion the “decomposition” could be representedsimply as a first-order reaction, where the concentration, X,of the decompositionproduct would be given by

X = XW[l-exp(-qt)]

(10,l)

(10.2)

or ln[l-

= -qt

; U ,

where q is a constant, t is time, and XW is found by iteration. This allows (from accelerated studies) a visual estimate of the worst appearance that a product could take on. The appearance of tablets can be treated differently and will be discussed later.

In ~arenteral solutions, physical stability includes interaction with a container and changes in chemical com~osition that give rise to physical changes. The latter will be discussed first. ne manifestation is slight discoloration. Thiamine hydrochloride solutions, for instance, may discolor slightly without showing detectable changes in content of parent compound. Such discolorations can be followed as described immediately

Antioxidant Acetylcysteine Ascorbic A,propyl and gallate Citric acid (chelator) edetate Sodium (chelator) Sulfites Thioglycerol Thiourea Tochop~erols

0.02-1% 0.005-0.02~ ‘Variable* 0.01-0.075 0.1-0.15% 0.1-1 .O% 0.5-1 .O 0.05-4.075

*Citric acid can be present in large amounts if it is present as a buffer (as well as present as a chelator). Source: Table constructed from data published by ende en hall (1984).

above, and at times they are detectable analytically. They are often oxidative in nature and metal ion catalyzed. Such a case in captopril (Lee and Notari, 1977). has reviewed the stability aspects of parenteral products oloration is often either photochemical or oxidative. has summari~edthe usually usedantioxidants and chelating agents. These are shown in Table 1.

ften a pare~teralsolution will develop a swirly precipitate upon storage. This is lent in vials and is usually an interaction with either the glass or the may be difficult for the uninitiated to detect such slight changes, and rson to use for this type of evaluation is a parenteral inspector. Itis difficult to estimate the extent of the precipitate; it can be done by mechanical counting (e.g., with a Coulter counter), but the results are diffi to interpret,Often the count pond to the “severity of the swirl.” re to the point is how many S type of stability, then the a box of e.g. 144 vialsisplaced o vials can be examined from time to time, and one may establish how many vials have become swirly.This number can then be treated in proper fashion to evaluate the severity of the problem, i.e., the stability parameter would be the number of d be no swirls at all in the preparation, and if aken (which is wise),then an improved product would stability program will establish the percentage proba swirl at the end of the expiration period. At times it is necessary to lyophilize products that are chemically stable, simply because blem of swirls cannot be solved. S mentioned, the occurrence of swirls is usually a container interaction, and a change in the stopper or the glass may often eliminate the problem. Vials should always bestored (a) upright, (b) on the side, and (c) upside downto check the interaction with the stopper. In this way primary evidence can be established as to the culpability of the closure.

McVean et al. (1972) reported on the case of a parenteral solution (morphine) where ‘cwhi~ker~” occurred at the tip of the ampul in a large percentage of ampuls upon room-temperature storage. This isa defect that will occasionally occurin a product. It is due to pinholes in the glass. The solution wicks out, and the liquid evaporates on the outside. The solid that is formed serves to wick out more solution, and long crystals or “whiskers” may occur. One might ask why the pinholes have not been detected in the dye test used for autoclaved ampuls. There are two reasons. One isthat the hole may betoo small for detection (about 0.5pm is the detection.limit). The other is that the ampul was tight at the time of manufacture, but the heat sealing line was run too rapidly, or the flame temperature was incorrect, so that the glass did not have time to anneal properly, and the strain caused the crack during storage (not immediately after manufacture).

Sometimes a cloud willappear in a product as the storage time progresses,and this is most often due to chemical changes in the system. If for instance an ester (e.g., polysorbate, which is a fatty acid ester) hydrolyzes, then the produced acid may be poorly soluble. If the solubility is denoted S, then the following holds: If the reaction in general is written A+H20+

(10.3)

where A is a drug of initial concentration A0 and is the decompositionproduct with order, the concentration ility S (which is assumedto be limited). Assuming first [B](10.4) = &[l -exp(-kt)] At time t* the solubility will be exceeded, and t* is what is denoted the cloud time. t* is given by S= (10.5) &[l

-exp(-kt*)]

[

In 1 --= -kt* If

(10.6)

>>S then this simplifies to S t” = kA0

(10.7)

2.8

3.0

3.2

1o o o n

3.4

Cloud times of a parenteral diluent containing polysorbate 80 and maleic acid. (Constructed from data published by Carstensen, 1972.)

Taking logarithms gives

ln[t*] = -1n[A0] In -

= -I~[Ao]+Q---

AA H H -- EE

(10.8) --~ RT

+e"

are constants. This shows that the cloud times can be plotted by where Q and Arrhenius plotting. Such plotting is quite predictive, as is shown in Fig. 1. The precipitation may also occur by the solubility product being exceeded, or from any situation leading to a product with limited solubility. There are other causes for precipitation on storage, one being the original useof a metastable form, so that the solutions in question, in fact, are supersaturated solutions. It was the author's experience, at his tenure at Hoffmann-la Roche in 1965, that a product to be introduced (Taractan Injectable) was in this category. Severalpilot batches had beensuccessfully made, but the first production batches precipitated, a more stable polymorph crystallizing out, This necessitated reformulation to a lower strength (corresponding to the lower solubility of the stabler polymorph) and subsequent resubmission of datato the FDA. This points out the importance of careful preforrnulation studies of the solubility of compounds. Errors of the above type are costly, both in terms of resubmission and in lost market time. Even official products fall into this category. Calcium gluceptate is used to treat calcium deficiency and (USP XX, 1980) is highly water soluble (up to 85%). Solutions, however, showa tendency to precipitate on standing at room temperature (Muller et al., 1979). The storage time required for precipitation is a function of the commercialsource, as is pointed out by Suryanarayanan and Mitchell(1981). It wasshown that the precipitate was a

sparingly soluble crystalline hydrate, and that the raw material was an amorphous (much more solu le) form of the drug. Seeds, and unfortuitous ratios of alpha a epimers of the calcium gluceptate, catalyzed the precipitation. n and crystal growth phenomenon c cars tens en and ornedo and Carstensen, 1985), and as such it can be impaired or prevented by inhibitors. se are often viscosity-impairi ,and hence the stability of the stances (carboxymethyl cellulose for insta component becomes important. The loss of this can be detected by following viscosity. The viscosity of these agents is often ingham bodies, i.e., they value. The correct wayof checking the is, therefore, withe.g. viscometer, so that a rheogram can be drawn. In this fashion it is possible to check both changes in yield valueand slope of the rheogram (apparent viscosity). For very ~ u i dsolutions (dilute aqueous solutions) this is difficult, and most often it is best o pi~ettes(with different edby the useof an stwald-Fens~~ pipe should be used in iscase,because the nce in the measuredvisc easure of the yield value (although calculation of the yield value from the difference is a priori not yield value and apparent viscosity are functions of concentratio t al., 1980); in a multicomponent s y s t e ~ there will usually be one onent responsible for viscosity, and it is the brea~downof this one compound that would be of importance. ften when drastic S occur in viscosity, bacterial contaminatio~ can be sus~ected. recipitation is tied into solubility, as seen in the foregoing. ~olubilitycan be augmented by various means. In the case of cloud times,the use of cosolvents (e polyethy~eneglycol)willincrease the value of S. her methods are the use o proach and the use of ~omplexation. ehdizadeh and Grant (198 ) on the complexation behavior of griseofulvinwith fatty acids. Order of ~ a g n i t u d eincreases in solubilitywere reported.

The main types of changes in ~ppearanceof oral solutions (syru loss of dye, precipitation, and bacterial growth. Precipitation ha with to somedegree, but somecases particular t solutions willbe mentioned. Change indye content willbe treated below. rial growth willbe treated separately. e dye in a vitamin syrup, a Scott et al. (1960) showed the loss that it could be treated exactly like a substance. re dictions by gradation in solution, b are quite goodin the case homogeneity is good. igure 2 shows an example of this.

isperse systemsare suspensions and emulsions. The rationale for the carried out on these will be discussed below.

2.9 3.5 3.4 3.3 3.2 3.1 3.0

100017” Arrhenius plot of FDC Blue Dye #2 in a syrup. (Graph constructed from data by Scott et al., 1960.)

It would be desirable to have a suspension that did not settle (and there are such suspensions), but the general rule isthat a suspension willsettle, and therefore there are two parameters that are followed in this respect, namely sedimentation rate and sedimentation volume. When the sedimentation volumes are small, then there is a tendency for the suspension to cake, and hence various types of shaking tests are carried out. Tests can be purely subjective, in that a tester notes that e.g. the §uspension after three months’ storage at 25°C was “difficult to resuspend, leaving some cake at the bottom,” Such subjective tests should always be included in a pro but more quantitative means are desirable also. A. typical quantitative test is to rotate the bottle under reproducible conditions. The type of setup used for solubility determinations is a good type apparatus for this purpose. The bottle is rotated x rotations, a sample of the supernatant is taken, and it is assayed. (This assay need not be stability indicating.) This is then repeated for twice the number of rotations, four times the number of rotations, and eighttimes the number of rotations. The time-relation of the assays is similar to that of a dissolution curve (although the phenomenon is redispersion), and it can often be represented by 0.9)

l’== (1 Yw[l -exp(-kt)] Yw,the asymptote value (found by iteration), should equal the dose, if caking has not occurred. The value of k is best found from the logarithmic presentation mode:

[ ~]

In 1”-

=-kt

(10.10)

Sodium

Lauryl

Sulfate

(rng/mL)

~esuspensioncharacteristics by ~on~rolled rotation. (Graph constructed from data published by Lemberger, 1967.)

k and I”& can then be found by data treatment for extrapolated values or assessed in a room-temperature stability program to estimate the stability of the resuspendability parameter. Suspendabilityis also improved by the use of surfactants. Figure 3 shows a suspension isotherm (Moore and Lemberger, 1963) of the zinc oxide/sodium lauryl sulfate system. Such suspension isotherms should be carried out prior to the formulation of suspensions. Theyare in general not carried out in the prefomulation effort, but rather by the formulator. One way of accelerating the settling is to place the suspension product on a shaker at e.g. 37°C. This makes particle movement more rapid and allows the fine particles to slip into the interstices of the larger particles, hence promoting a close packing. Thiscan then be used to judge qualitatively whether caking will take place. It might be thought that centrifugation would be a good way in which to ‘‘ac~elerate’~ sedimentation, and the Stokes law indeed predicts this. However, it gives only an acceleration of the “initial settling rate,” and the further settling, and the caking phenomena in which the formulator is interested, are not well predicted by this method. Some caking is due to crystal growth, and this is accelerated by the use of freeze-thaw tests, i.e., alternating the temperature every 24 h from e.g. 25°C to -5°C (or some other low temperature above the freezing point of the product). The temperature cycle will promote crystal growth, and the effect of this on the product can be assessed. The freeze-thawcycle has the advantage of emulating (and overstating) some real conditions to which the product could be exposed during shipping. Zapata et al. (1984) have described the effect of freeze-thaw cycles on aluminium hydroxycarbonate and magnesium hydroxide gels. Coagulatio~after freeze-thawcyclesled to the formation of aggregates that werevisible.These aggregates were particles in a primary m i n i ~ u m and , these were only reseparable by ultrasonic treatment. The freeze-thaw cycle affected content uniformity of both the gels, but the treatment did not alter the surface characteristics or the morphology

(as judged by x-raypowder diffraction). t didcause a reduction in the acid neutralization rate, and the rate of sedimentation increased. The effect was pronounced after the first cycle (and indeed most of the effect occurred at this point). The duration of freezing was not important, but the aggregate size grew inversely with the rate of freezing. The use of polymers in the suspensions reducedthe effects of the freeze-thaw cycle. Freeze-thaw cycles (aside from being a stability monitoring tool) can be used to screen products as well, the best of a series of suspensions or emulsions being the one that stands up best to the test. This on the surface may belogical, but without a theoretical basis it is difficult to judge the generality of such a statement.

If a suspension is particulate, then the particles will (approximately) settle by a Stokes law relation, i.e., the terminal velocity, v, is given by (10.11) where the constant g is gravitational acceleration, Ap is the difference in density between solid and liquid, q is the viscosity of the liquid, and d is the diameter of the particle. The final apparent volume of the sediment, provided it is monodisperse, would be given bythe fact that in cubical loose packinga sphere of diameter d will occupy the space of its confining cube, i.e., the sedimentation volume will be V=n*d3

(10.12)

where n is the number of particles per cm3 of suspension. Since their density is p g/cm3, then (denoting the dosage level Q g/cm3) the following holds:

p n n;d3

(10.13)

so that, solving for n, y1=-12.6 pnd3

(10.14)

which inserted in Eq. (10.12) gives (10.15) In this view, each particle touches its neighbors. The potential diagram from two particles is as shown in Fig. 4. When the particles touch, the potential energybecomesexceedinglylarge (x =0), and from an equilibrium point of view they will be trapped in the primary minimum, which isthe deep minimum at short distance in Fig. 4. Hence it becomes difficult to separate them, and the precipitate becomes a cake. This would prevent redispersion by shaking and would make proper dispensing impossible. It is a formulation goal to prevent this from happening, and this is done by adjusting the

-20

Potential energy diagram for two particles.

zeta potential, as will be discussed shortly. From a f o ~ u l a t i o npoint of view, it is better to have the particles at larger distances, e.g.,in the secondary minimum occurring at longer distances (Fig. 4). A discussion of the connection between caking tendency and the so-called zeta potential is beyond the scope of this book. Suffice it to state the following: particles are suspended in a liquid, they acquire a charge (and the liquid acquires a similar opposite charge, to maintain electroneutrality). The zeta potential is related to this charge, and caking is prone to happen if the charge potential is outside a range of -10 mV to +l0 mV. If the zeta potential is highit can be lowered by the addition of negatively charged ions. Highly valent ions (e.g., citrate) are preferable. On the other hand, if the zeta potential is low, then it can be increased by the addition of positively charged ions (e.g., aluminium ions). The zeta potential is measured with a zetameter, In this the particles are placed in an electrical field (betweentwo electrodes, the voltage of which can be adjusted), they are tracked under a microscope, and their velocity isdetermined. The relation of velocity to voltage allows determination of the zeta potential. It is worthwhile occasionallyto check the zeta potential in a stability check of suspension (and emulsion) products. Counterions could be adsorbed and hence lose their capability of keeping the zeta potential close to zero, and this, in turn, could son for subsequent caking. en. the zeta potential is closeto zero, the suspension will be flocculated, i.e., the particles are positioned in the secondary minimum. The floccules are large and hence settle more slowly, but on the other hand the sedimentation volume is large. Since the particles are in the shallow minimum (small potential, i.e., easyto disrupt), they are easy to resuspend. There are suspensions that do not settle. Here the yield valueof the suspension is so large that the gravitational force doesnot exceed it. In this case it is veryimportant tocarry out complete rheological profilesat different time points in the stability program, to insure that the yield value is not changing. In such a system the yield value (Carstensen, 1973) is a function of the solids content and the viscosity of

the medium. If the viscosity imparting substance deteriorates, or if the flocculation characteristic (the “diameter” of the particles) changes, then the yield value may change, and what originally was not prone to cake might at a later time have such a propensity. It has been stated elsewhere that for ingham bodies, a yield diameter of the bottle can be calculated and below this bottle diameter there will be no settling.

The rational treatment of sedimentation rates has been described by Carstensen and Su (1970). Since the suspension, when placed on stability, has just been well agitated, the floccule size is not the same as it will be at equilibrium (it will be he first part of a settling curve is, therefore, governed by the reforming of the equilibrium floccule, and the latter part is governed by settling towards the equilibrium sedimentation volume. A. typical plot of the final settling phase of kaolin suspensions is shown inFig. 5. The intercept does not correspond to full height, because the settling is the final phase. The first phase, as mentioned, consists of reflocculation of the e q u i l i b r i ~floccule (which does not exist at time zero, because the suspension has been thoroughly shaken at that point). The sedimentation curve is, therefore, two-phasic, and the equation for the settling curve is Y -Y, = A0 exp(-kot)

1 exp(--kl t )

(10.16)

and the curve can be deconvoluted by feathering, or by programmed four~parameter techniques.

ethyl, ethyl, propyl, and butyl esters of 4“ydroxybenzoic acid are used invar mbinations in antacid suspension (and other pharmaceutical) products.

Time (hours)

100

120

Settling of kaolin.suspensions. (Constructed from data published by Carstensen and Su, 1969.)

51-

40 60 Time (min)

The least squares fit equation is y =4.69 -0 . 0 6 6 ~with R2 =0.99. (Graph constructed from data published by Cardenas et al., 1994.)

antacids havehigh pH values, and hencehydrolysis of the estersoccurs. The rationale for using several in combination is, exactly, to allow a certain amount to remain to retain preservative qualities of the suspension. An assay of the four esters and the parent acid (one of the decomposition products) in products where all occur has been described by Schieffer et al. (1984).

The 1987 Guidelines require testing of suspensions for dissolution. Cardenas et al. (1994)havedescribed the dissolution profiles from suspensions ofbenzoyl metronidazole, and a graph constructed from their published data is shown inFig. 6. The curves should, by all rights, follow a cube root law without lag time, but they do not do so. If adjusted for amount not dissolved at the end (in the figure, 10%) they will adhere to a sigma minus plot.

A suspension is,as the name implies,a two-component system consistingof a solid and a liquid phase. (Gas phases are considered nonessential in this connection). Obviously, the solubility of the compound is a function of the temperature, and at a given temperature above 25°C this solubility will be reached. Testing about this temperature obviously has no meaning as far as suspension stability (neither physical nor chemical) is concerned. Prior to starting a program, this temperature should be established, so that unnecessary sampling stations can be avoided. .7.

Some semisolid systems (ointments and suppositories) are suspensions. Their testing is not different, in general philosophy,from what is describedabove, except that the rheology is checked differently. Davis (1987) has reviewed so~histicatedmeans of checking the stability of such systems.

The factors checked for in stability programs of such products are the following: 1.Consistency,fell 2. Viscosity 3. Polymorphism

to the touch

It is mentioned elsewhere that migration of a “disperse” phase within a semisolid product is quite possible when another phase is present. This situation may occur in the case of the use of benzocaine in, for instance, a suppository wrapped in aluminumfoil coated withpolyethylene.Polyethylenelining of aluminum wraps of suppositories is used to prevent contact between the metal and the suppository, and in most cases this has a positive effect. However, a partitioning of drug or additive between the two phases may be possible if the drug or additive issuspended in the suppository. Denoting its solubility in the polyethylene S, and the solubility in the suppository base Ss, the compound would disappear from solution in the suppository at a rate proportional to Sp-Ss,and “disappeared” compound would be replenishedby dissolution from the solid phase. The rate of disappearance would be governed in that the value of Spwould increase by a sigma minus relation (i.e., in the same manner as the appearance of decomposition product in a first-order reaction), and this then would be the overall “loss” of compound as a function of time. Since this is a first-order overall relationship, the “decomposition” would, initially, appear to be first order.

Polymorphism can be followed by x-ray analysis and in some cases by thermal methods. There is, in fat systems, the possibility of trans esterification, and this can be tested for chemically. The problem of morphology changes isoften of particular importance and of particular frequency in the case of suppositories. In this type of product, it is also important to check for migration of suspended/dissolved substances. Often a substance is added to a suppository as a suspended particle, which is soluble in the suppository base to some extent. The phenomenon of dissolution will, of course, become evident by checking the particle size as a function of time. If a substance is soluble in the base, then it is preferable (if possible) to saturate the base with it at the onset. For this reason it is necessaryto determine the solubility (S gm/gm) of the drug (or other) substance in the base. A Van’t Hoff plot [solubility as a function of temperature (T’K), i.e., plotting l n [ q versus 1/Uwill allow extrapolation to room temperature. In manufacturing it is advisableto dissolve the drug (or other substance) to the extent of its solubility during the intermediate temperature phase of manufacturing (where the preparation isstill quite fluid) and then suspend the rest at a lower temperature. An example is ascorbic acid, whichis a good antioxidant in Carbowax bases. To exert its antioxidant action it must, however, be dissolved (and it is quite soluble in polyethylene glycols). Dissolved drug (or other substance, e.g., benzocaine) will diffuse in the suppository base, and can, for instance, partition into polyethylene linings of the suppository wrap.

elease rates are important in many topical preparationss in particular in transdermal preparations. Here there are several investigational methods available. In-vitro methods involveplacing the ointment on a membrane and ~ e a s u r i n g the appearance of drug in a receptor compartment on the sink sideof the membrane. oelgaard and Mollgaard (1983) have, for instance, described the in-vitro release of linoleic acidthrough an in-vitro membrane. They mounted abdominal human skin in one case and skin from hairless rats in another to open diffusion cells.The dermal side was bathed with a rece tor medium stirred at 37°C. The medium was 75 mL, of 0.05 N phosphate buffer ”-7.4) which contained 0.05% ic F68 and 0.01% butylhdroxytol~ene, atter two ingredients added in to increase the lipid solubility. Linear, Ficksian diffusion curves were obtained. In a stability program, such tests are obviouslyuseful and should be repeated periodically, but an “internal standard” or “calibrator” should beusei.e., a stable test su~stance, the diffusion of which is known (e.g., salicylic acid). ther pseudo-in-vivo methods involve shavedor hairless rabbits, or cadaver skin. The interaction between ointment and container (patch) should also be part of the stability program. Some ofthe testing applicable to semisolid emulsion systems also is applicable to ointment systems and will be discussed at a later point.

An emulsion should be thought of as a metastable system. In most cases the emulsio~ system (Fig. 7) is thermodynamically more energetic than the ground state system, whichwouldsimplybe the totality of the twophases, separated, There will, therefore, always be the potential for oil droplets re-merging inan a t t e m ~to t create the thermodynamically stable system. Emulsion systems are taken orally ( L i p o ~ a n t r a c i n T ~oche), s parentally (as parenteral fat emulsions), and topically (creams). Isi

The factors that stabilize the emulsion systemare a layer of surfactant and protective colloid on the exterior of the droplet. The amount of these two must be such that they

eparate Phases Emulsion system.

Emulsion

cover the entire area of the droplets, otherwise coalescence will occurto the extent that the area, A, of the droplets will be reduced to such a point that it now will be completely covered by surfactant and protective colloid. If, for instance, 1 g of emulsion contained W g of droplets of a size d pm and the oil had a density of p g/cm3,then there would be n droplets per cm3,where n is given ach particle has a surface area of d 2 , so that the total area is (10.17) ~ x a ~ 10. ~ l.l e

If the density of the oil is 0.9 g/cm3,the amount of oil phase per cm3,is 0.75 g, and the diameter of the oil lobules is 10 pm (lod3cm) what is the surface area of the oil phase? Answer.

A=-.-."-0.75 6 -5 103cm2 10-3 0.9

(10.18)

a10.2, ~ ~

If a surface active agent of molecular weight800 and cross-sectional moleculararea of 30A is present in a concentration of 0.2% will that suffice to cover the surface in Example 10. l? Answer

2 mg/cm3 = 2/800 = 2.5 loe3 millimoles = 2.5 moles,which in turn equals 2.5 loM6. 6 = 1.5 1OI8 molecules = 30 1.5 10l8A2 = 4500cm2, this is the surface the surfactant could cover. This is slightly less than the 5000 cm2 surface area of the oil, so that the entire surface of the oil globules cannot be covered by the surfactant. *

The above calculations are oversimplified. They assume, for instance, that all the surfactant is adsorbed onto the oil, which is not the case. It is important, however, to check, originally, whether enough surface coverage of the oil is provided for. If not, there will be an initial shrinkage of surface area (increase in droplet size) attributable to this. ence, if the coverage of the droplets with surfactants and/or protective colloid is complete at the time of manufacture, then the droplets will grow in size as time progresses. owe(1965) for instance demonstrated that the globule size decreases withincreasing surfactant concentration, as shown in Fig. 8.

The breakage of suspensions will be dealt with shortly, but (Fig. 7) it might be suspected that breakage would be a function of Stoksian motion [Eq. (10.1l)], i.e., the globulesmove and collide and hencecoalesce.Thisis true in a sense, but

Sulfate

Mass Mean Diameter, pm

Effect of emulsifier concentration of globule size. (Graph constructed from data by Rowe, 1965.)

the conclusion that might be drawn would then be that to increase viscosity [Eq. (10.1l)] wouldreduce the severity of such impacts. However, Siragusa (1995) has demonstrated that although increasedviscosity to some extent makes an emulsion more stable, the more important factor is the stability of the surfactant/protective colloid systemat the interface. From a stability point of view, there is a correlation between the overall emulsion viscosity and the globule size. Figure 9 shows a typical exampleof viscosity as a function of droplet size and phase ratio (Sherman, 1964). Hence, checking viscosity in the stability program, in a manner of speaking checks the globule size, which is the prime indicator of potential for progressing creaming and breaking. The viscosity is usually checkedby a cup and bob method. The limitations of this willbediscussedin the sectiondealingwithsemisolid emulsion systems.

d = 3.3

0.2

0.4

0.6

0.8

Volume Fraction of Oil

1.o

Viscosity as a function of droplet size and phase ratio. (Constructed after data published by Davis, Sherman, 1964.)

ect measurement of the droplet size can be accomplished in several ways: py, electronic (Coulter) counters, photon correlation spectroscopy (for particles that are very small, Davis, 1967), diffuse reflectance spectroscopy (Akers Lach, 1976) and the measurement of ultrasound (Rassing and Atwood, 1983). is (1987) points to the importance of choosing the proper techniques. He cites an example where a fat emulsion was tested for stability (as regards droplet size and distribution). The accelerated test used was a shaking test. The tests used were (a) microscopy (large globules), (b)electronic counting (medium size globulecount) and (c) photon correlation spectroscopy (small particle count). Figure 10 shows the results.

1 = Coulter counter before 2 = Proton correlation spec. before 3 = Coulter counter after 4 = Proton correlatin spec. after

1 and 3 diameter before 2 and 4 diameter after

Particle size analysisof acceleratedtest of emulsion system. (Constructed from data published by Davis, 1987.)

It is obvious that the small globulecount does not change much, but that the intermediate count changes a lot. (The large globule count WO complement fashion to the intermediate count, and thiswas microscopy results.) What is important in this particular case was a parenteral fat emulsion) isthat there was formation of la visible to the naked eye), and that these could have had a bearing on the toxicity of the product. This demonstrates that one method in itself is not enough, and that not one but several methods should be considered. eng(1984) has advocated electrical conductivity as an overall, common means of d e t e ~ i the ~ ~state g of dispersion of an emulsion §ystem,and he shows that this parameter changes significantly overshort periods of time, if the emulsion system is not satisfactory.

let size, is chemical breakdown in the surfactant. Nonionic surfa~tantsare frequently used, and they esters that may hydrolyze or interact with other components of the emulsio t of the formulator's job is, in independent experiments, to determine the p profile and interaction potential of the surfactant (in a system simply c of the aqueous phase) with the other additives of the emulsion system. be done simply by cloud times (at accelerated temperat~res)if the acid o from the hydrolysis or the interaction product is poorly soluble (as it is in the case rbates and arlacels). e problem with nonionic surfactant hydrolysis is exactlythat it produces a fatty acid, which may become part of the oil phase and h e n c ~(aside from providinglesscoverageof the oil droplet) change the emulsion characteristics of the system. In general the formulator also d e t e ~ i n e sthe (hydrophilicIlyophilic balance) of the systemheworkswith and matche surfactant used [Atlas Chemical Company (now IC1 Americas), 19631. of the em~lsifiercan be adjusted by mixing two emulsifiers, e.g., arlacel 85 has an of 2.0 and polysorbate 80 one of 16.5. If an emulsion system requir of 10 for instance, then the ratio of polysorbate (x) to arlacel (1-x) would be given by

2(1 -X) 1 6 . 5 =~C: 10

(10.19)

8 x = - 14*5= 0.55

(10.20)

In emulsion formulation, the type of emulsion isof concern. If it is desiredto make an oil-in-water emulsion(olw,i.e., oil isthe discontinuous phase), thenit is important that phase inversion not occur. Investigating this possibility must be a task gram (and is usually carried out by the formulator, not the st often phase inversion isassociated with creaming and separ-

ion T e m ~ e r ~ t (* ur~ Coalescence rate versus inversion temperature.(Graph constructed fromdata published by Enever, 1976.)

ation and will benoticed inthe appearance testing of the emulsion. Suchphenomena lead to graininess of feel. In so e cases part of an emulsion willinvert, another not, and then there is a distinct difference in appearances in various regions of the the possibility for inversion should always be considered. It is the more likely the closer the system isto a close-packed systemof spheres. In this connection, ther of the formulator9stasks should be to determine the inversion temperature. is is at times used to advantage in the manufacturing step, in that, in producing the emulsion, the inverse emulsion is produced at high temperature; this is then cooled, and at the inversion temperature, the “correct9’type will result. onv version in this manner gives rise to very small globules, and homogenization is then often f an inversion temperature exists, then accelerated testing above gless. So preliminary testing is alwaysadvocated, if accelted, the philosophy beingthat there is no sense in testing a system above a temperature where it converts to a physical state that differs from that at room temperature (or recom~endedstorage temperature). Enever (1976) has shown that there is a corr ation between phase inversion temperature and the rate of coalescence (Fig. 11). is possible to use a combination of sedimentation ow fractionation and photon correlation spectroscopy to record droplet sizes in fat emulsions, and this would appear to be an excellent technique for studying the coalescence of finer spheres, and hence to obtain an extrapolatory tool early on in the storage of an emulsion system. *

It has been mentionedthat there is a gross correlation between viscosityand globule owever, the rheological characteristics of an emulsion system in general S on other factors as well (Sherman, 1955): he viscosity of the internal phase

2. 3. 4. 5. 6.

The viscosity of the external phase The phase volume ratio What emulsifiers are used and in what amount The electroviscous effect Distribution of particle sizes

The appearance of the emulsion will be a function of globule size, and Table 2 gives a gross correlation of thesetwo factors. When an emulsion breaks, the hyponatant, rather than being a solution, will haveone of the two firstappearances in the table, i.e., will also be an emulsion, but with very fine droplets.

It can be concluded from what has been mentioned that the reasons for breaking would include 1. Chemical incompatibility between the emulsifier and another ingredient in the emulsion system (Borax and gum acacia is a case in point) 2. Improper choice of surfactant pair (e.g., wrong 3. High electrolyte concentration 4. Instability of an emulsifier 5. Too low a viscosity 6. Temperature As shownin the foregoing, breaking and creaming of emulsions are the typical defective criteria to be looked for in stability programs. Breaking implies that the emulsion separates into two distinct phases (Fig. 7). If this is a slow process,it often manifests itself in the appearance of small amounts of oil particles on the surface, and it then is referred to as oiling. When separation into two emulsions occurs (as described above), then the phenomenon is calledcreaming. A rapid test for this is to dip a finger into the preparation and notice if there are different “colors” present rown, 1953). Also, a creamed olw emulsion will not drain off the skin with ease, and the converse holds for a creamed wlo emulsion. A few words regarding the effect of ionic substances and the actual process of flocculation and coalescence are in order. Van denTempe1 (1953)demonstrated that flocculation and coalescence are two different processes. ~locculationdepends on electrostatic repulsion (and is akin to the zeta-potential considerations discussed previously). Coalescence depends on the properties of the interfacial film. Correlation Between Globule Size and Appearance of Emulsions Globule size (pm) >0.005 0.005-0.1 0.1-1 >l

Appearance

Translucent (trans~arent) Semitransparent, gray ~luish-whiteemulsion emulsion Milky-white

Cations, as a whole, are less soluble in the oil phase than anions, and this gives rise to negatively charged droplets (akin to the creation of a zeta-potential in suspensions). The potential drop over the film depends on the nature of the electrolyte (and it should be noticed that there is a diffuse double layer in both liquids as opposed to the case of suspensions, wherethere is onlyone diffuse double layer). Electrolytes may either improve or worsen the stability: If they eliminate the protection offered by the surfactant /protective colloid system then coalescence ost often electrolytes have the effect of reducing the emulsifying powers of surfactants and causing salting out or actually precipitating the surfactant. However, in some cases, electrolytes will favorably affect the potential drop over the two double layers, and in this case they may stabilize the suspension system. is

Semisolid emulsions (cold creams, vanishing creams) are not different, in general philosophy, from the above, except that the rheology is checked differently. Davis (1984) has reviewed sophisticated means of checking the stability of these types of systems. He lists the following properties as being important in stability programs for semisolid emulsions: 1. Particle size 2. Polymorphic/ hydration/solvation states 3. Sedimentation/creaming 4. Caking/coalescence Consistency 5. 6. Drug release Of these, particle size, sedime~tation/creaming,caking-coalescence, and consistency have been discussed earlier. Following viscosity as a function of time is here of particular interest, The problem is howto measure the viscosity, and what viscosityin essence means. (1987) points out that changes in viscoelastic properties are much more sensitive than simple continuous shear measurements (Barry, 1974). He demonstrates this via data published by Eccleston (1976).Here (Fig. 12) the variation of the dynamic viscosity (q) and the storage modulus (4) are shown and compared with the same type of graph for apparent viscosity (p’) from continuous shear experiments. It is obvious that the two former measurements are much more sensitive.

The most important concern about transdermals is the release of drug substance from them and the stability of this property. Other properties (stickiness, appearance, etc.) are of importance as well, but the release characteristic is paramount. Kokobo et al. (1991) have described a means of checking this in vivo by using a single diffusion cell. The volume could be, for instance, 2.5 mL, and in contact the diffusion area could be of the order of 1cm2.The matrix is placed, e.g., with a 40% polyethylene glycol solution, which can be, e.g., removed in 500 pL quantities.

12 10

4 2

0 0

lime (days)

-2

0.3 0.2

0.0

Days

(a) Dynamic viscosity(v) and storage modulus (4)and (b) apparent viscosity (p’) as a function of storage time for cosmetic creamsmade fromstearyl alcohol.(Graph constructed from data published by Eccleston, 1976.)

obo et al. (1994)have reported on the interaction betweenpreitive adhesives and drug combinations used in transdermals. Their data are shown inFig. 13. The data fit neither a diffusion equation (In of retained versus time) nor a square root equation directly. It would appear that if one allows for either an initial dumping in the diffusion equation (or includes more than one term arrer equation) or a lag time in a square root equation, then the data will

0 0

Time (hrs) Release of dipropylphthalate from 2-ethylhexylacrylate acrylic acid copolymer (2EHA/AA). (Graph constructed from data reported by Kokobo et al., 1994.)

fit either. A modified diffusion equation is probably the most likely. The authors suggest the useof the illiams- an del-Ferry equation for fitting: {log D}-A =

-896

51.6

+ (T-Tg)

(10.21)

where D is the diffusion coefficient, A is a constant, and Tgis the glass transition temperature of the polymer.

Accelerated testing of physicalproperties of disperse systems not is as clear-cut as for instance chemical kinetics prediction. For instance, the stability of properties of semisolid materials is very difficult,for instance, for creams and ointments that give rise to bleeding there does not seem to be any reliable predictivetest. Yet a series of stress tests are used for disperse systems. They include Shaking tests ~entrifugaltests Freeze-thaw tests Elevated temperature tests It should be cautioned that although these types of tests can be performed on a comparative basis (Is formula A better in onerespect than formula interpretation, other than saying that A is better (or worse) than B, is and predictiveaspects are somewhatlacking,because the phenomenatestedin the accelerated tests do ot necessarily mimicwhat will happen in room temperature storage a/o shipping ( avis, 1987; Rhodes, 1979a and b).

For the freeze-thaw test, the question is whatthe minimum temperature should be, temperatures from -5” to +5”Cbeing the most common. -5°C frequently gives rise to phase separation and irreversible changes that would not be seen in usual temperature ranges (Nakamura and Okada, 1976),but again, such tests may be used to select a “presumably best” formula from a series of preparations in product development. Results of a typical freeze-thaw cycle are shown in Fig. 10. ~entrifugationhas been usedby some investigators (Tingstad, 1964; Hahn and ndracek et al., 1985). The general idea is that g can be increased city predicted by Stokes’s law (Eq. 1l), but often the stresses caused by centrifugation may cause coalescence, which would not occur during normal collision stress. Some investigators claim fair successin predictions by this means, but as avis (1987) cautiously states, “as a general rule it can be stated that systems that accelerated stress conditions should be stable under normal storage conowever the corollary is not necessarily true.” That is, if the preparation fails the test it may still be all right, but if it passes the test it should be all right. ~ l t h o u g hthis may be true overall,one can visualize that if a preparation is centrifuged right after manufacture, then the stress does not include the chemical changes (surfactant decomposition for instance) that occur on storage, and in this it may give too optimistic a prediction. uscall et al. (1979) have measured phase separation at several different centrifugal gs and have established from these data a so-called coalescence pressure. This (again recalling that the test does not account for chemical changeson storage) may be an appropriate parameter. Onepredictive method in formulation is the correlation afforded by coalescence rates (Fig. 1l), and this is rational in selecting the “best,, of many formulations; in general the system with the highest phase inversion temperature is the best. The (nonchemical stability dictated) coalescence rate could theoretically be calculated prior to storage, and the difference between observedand calculated then attributed to chemical stability causes. For emulsions, it should again be pointed out t t rapid creaming and necessarily mean rapid coalescence. r emulsions there have tie zeta-potentials to blhau et al., 1977) that attempted emulsion behavior on storage, but the generality of such an approach has been test is usuallycarried out at 2-3 hertz ( avis, 1987), and the philosophy here is to intensify the collision frequency between globules (and to some degree also the intensity). This is therefore considered an accelerated test, but it actually is part of the product life (transportation). In any event, it should be included in protocols and simply reported.

ciarra has reviewed pharmaceutical and cosmetic aerosols (1974). Aerosols are solutions, primary emulsions, or suspensions (i.e., suspensionsin a suitable solvent such as ethanol) of active principle in chlorinated hydrocarbons, contained in a pressure can. Either a dip tube or a metering device connectsthe pressurized liquid contents to the valve. Upon activation of this, the internal pressurewillforce

the liquid through the valve orifice and atomize the suspension. The chlorinated hydrocarbon and the primary emulsion or suspension vehicle will evaporate, and the drug, in finely divided form, will be administered to the location of treatment (lung, skin). In general the physical instability of aerosols can lead to changes in (a) total drug delivered per dose or (b) total number of doses that may be obtained from the container. It is intuitively obviousthat the particle size range must be fine (i.e., the particles will have to pass through the valve). In general the primary disperse system is filledinto a seamless aerosol can, the valve assembly is attached, and the halogenated hydrocarbon is filled by pressure through the valve. The under-the-cap filling method has been described by Boegli et al. (1969). The halogenated hydrocarbon can, alternatively, be “liquid filled” at low temperature. For products that are moisture sensitive, this presents the problem of condensed ice and water in the product. As far as “cleanliness of operation,” aerosol lines are usually kept separate from conventional filling lines(Sciarra, 1974) (or the product is contract filled). Some attempts have been made to use ethylene oxide sterilization of the can (Joyner, 1969a, 1969b), and aseptic fillings (Harris, 1968; Sciarra, 1967) can be carried out.

Some testing methods are official in the USP (XXI). The Chemical Testin Manufacturers Association has developed a series of tests described in the ASC book (Aerosol Guide, 1981). Several testmethods are used to detect physical aerosol instability, viz., (1)unit spray content, (2) color and odor, (3) rate of leakage,(4) moisture and trace catalytical substances, (5) particle size distribution, (6) spray characteristics, (7) moisture and trace catalytical substances, (8) pH, (9) delivery rate, (10) microbial limit tests, and (1 1) container compatibility. Of the above, leak testing isofficialin the USP (XXI). Thisconsists of obtaining the weight loss after at least 3 days of storage and converting it to loss per year. If plastic-coated glasscontainers are used, the test should be done stant humidity. A faster method is to use an eudiometer tube described in the aerosol guide. This has the advantage of speed and also is advantageous in that it distinguishes between leakage from crimp versus leakage from valve gaskets. For spray characteristics a qualitative measurement isto spray onto paper that is treated with a mixture of dye and talc, as described in the CS A Aerosol Guide. also radiotracer techniques (Smith et al., 1984) and TLC graphic technjamin et al. 1983). The Aerosol Guide, p. 77 also describes a method whereby the spray is sprayed through a pie shaped wedge onto a rotator. Particle size analysis is the most important characteristic and hence the most important aerosol stability test. Sciarra states that particle sizes are between 1 and 10 pm and mostly between 3 and 5 pm. Particle size affects stability of delivery rate, effective dose, mass of drug delivered and of course the stability of the suspensionitself. The methods used are microscopy, sedimentation methods, light scattering, cascade impactors, and liquid impingers. Ifthe particle size distributions are determined by electronic methods (e.g., Coulter counter, Malvern), then allowance for solubility should be made.

et al. (1969)haveshown that the spray particle sizeisreduced by drug particle size, by concentration of drug, and by the valve orifice size. igher propellant temperature, vapor pressure, and using a surfactant in the formula also made the spray particle size smaller. Particle sizes are important for reasons other than physical stability. For inhalation aerosols, for example, it should be recalledthat particles larger than 20 pmdo not go past the terminal bronchioles, and particles at 6pm do not reach the lower alveolar ducts. Particles 0.5 to 5 pm reachthe alveolar walls and are intermixed with alveolar fluid (Idson, 1970). The chance for a 1pm particle depositing is less than 50%. There is therefore, particularly for inhalation aerosols, a very narrow particle size range of effectiveness. oisture testingisof importance, and except for foams, pharmaceutical aerosols are nonaqueous suspensions. Devices exist that willallow the transfer of the content of the can directly into a Karl Fisher apparatus. This is preferable over transfer by cutting the can open, since this method would allow for condensati into the product (which is chilled at the time of the opening of the cription of the can device for piercing the can is to be found in the erosol Guide; it allows direct sampling from the content of the aerosol. The moisture is measuredby Karl Fisher titration, and there are a number of commercially available instruments that can accomplish this. Pressure testing is also an official USPXXI method. A re pressurized gauge is d on the valve stem, and the valve isactuated so t it is all the way open. In the A Aerosol Guide, pressure testing is described. e method employs piercing the other tests directly through the valve. icrobial limits are described in the USP, e.g., betamethasone valerate topical The ~ i c r o b i alimits l must meet the requirements of the test c u s and ~ ~ e u ~ o ~aeru~inosa o n a s under the of ~ t a ~ h y ~ o c o caureus Tests to elivery rate isofficial in the USP XXL. The aerosol isallowed temperature-equilibrate at 25°C. The weight is determined, the can is then actuated for 5 seconds, the weight is determinedagain, and the delivery rate is thencalculated by difference.Delivery rates usuallychange on storage because of changesin elastomer hardness and gasket swelling. An apparatus is available from Peterson/Puritan Inc. that is accurate to four significant figures. In this assembly a solenoid can hold and actuate, and measure to 0.001seconds by stop clock (~ohnson,1972). Poiseuille's law appliesto aerosol spray delivery rates: Fisher and Sheth (1973) have shownthat delivery rate is linearlyrelated to the container pressure and that it is inversely proportional to viscosity of the can content. Also for a satisfactory system, the delivery rate will not to any great extent be a function of how much of the can has been emptied out. Of course if a can is emptied in one fell swoop, then the cooling effect of expansion may slow down the rate. Also, fractionation of propellant mix occurs and may lead to increased variation of delivery rate. a 'Valve testing and evaluation should always be done on the final f o r ~ u l (i.e., not on selected solvent systems). A pure solvent will not fractionate, and hence iation of spray rate may be smaller than with the final formula. inally there is the question whether there is an interaction between the can and the product. Can interaction and moisture content are closely related, since under

adverse conditions, the halogenated hydrocarbon will react with water and form a halogen acid that may corrode the can. Coating of cans can slow down the rate of this corrosion but not necessarily eliminateit. The control of moisture is therefore important not only for this reason but often also for the reason of chemicalstability of the drug.

These are mentioned here in distinction from aerosols; theyare mostly nasal spra S. In testing these, the droplet size is important in metered-dose sprays, since S droplets can reach bronchi and alveoli, which would be undesira~le,e.g., for delivery of corticosteroid treatment of rhinal disease. Yu et al. (1984) have described experimental setup used for determining the droplet size of flunisolid nasal lass chamber with an air inlet and a plastic stopper that has a hole that of the spray unit his is connected via a conical cavity to a cascade impactor and appropriate an This can be done (more expensively) by laser to followpossible holography (Uu et a1 h instrumentation may be used chan es in droplet size distribution as a function of time. ron (1990) emphasizethat the size of the particles rs in the efficiency of deposition of solids from called for a sampling chamber size of 500mL 94) have modified the Anderson Impactor as at the chamber volume greatly affects the perthe effect ofthe chamber ndersenSampler Inc., ufacturer recommends that, at a flow rate ) are 9, 5.8, 4.7, 3.3, 2.1, 1.1, 0.7, and 0.4pm for stages 0 to 7.

harmaceutical powders are for reconstitution into either suspensions or solutions. A prescription example of the former is chloramphenicol palmitate, where the arried out by the pharmacist prior to dispensing. An example tamucil, where the customer reconstitutes the product (e.g., in ples of solutions are Achromycin fM (which is a parenteral the-counter examples of oral solutions of this type are older produ gna Granules (LederleTM). Analogies in the food area are fruit hich are sold in packets and reconstituted by the c o n s u ~ e rto a certain volume. The main physical concerns in this type of product are appearance, organoleptic properties, and ease of reconstitution. nly the latter will be treated here. There are several reasons a powder may changedissolution time as a function of storage time. The most common reasons are (a) cohesion, (b) crystal growth, and (c) moisture sorption, which causesa l u ~ p i u~p gof powders. The latter is simply due to the dissolution and bridge-forming that occurs and is akin to what happens in wet granulation.

Chamber

Figure drawn from schematic published by VanOort et al. (1994).

60 0

1000 2 0 0 0 3000 4 0 0 0 5000 6 0 0 0

Chamber Volume (mL) Chamber volume vs. respirable dose. (Graph constructed from data published by VanOort et al., 1994.)

Cohesional force is the force between two particles, and cohesion in general is the stress (force percm2 of surface) that aparticle experiences due to the surrounding particles. Problems due to cohesion are particularly predominant when a powder is fine, and great fineness of a powderis often required for dissolution reasons. Cohesional forces are inversely proportional to the square of the distance between the particles, so that in storage, where vibration, for instance, may consolidate the powderbed,theseforcesbecome large, and the powder “cakes up.” This may give rise to problems in reconstitution.

There are two situations in c ~ y ~ t a Z g One ~ o ~istdue ~ .to the polymorphism. If the original product is either a metastable polymorph or amorphous, the conversion may occurin storage. For this to happen, some stress, e.g., the presence of moisture, must occur. The stress need not necessarily be moisture, conversion of a small amount of powder might occur in the filling head of the filling machine and then propagate in time. If the content of the drug substance is suchthat there are no neighboring drug particles, then this conversion is limited. Particularly, contact points allow for propagation of conversions in situations where the spontaneous nucleation probability is low.The presence of moisture will accelerate conversions of this type, once a seed of the stable polymorph (or in the amorphate situation, once a crystal) has formed. Crystal growth is,perse, not to be expected. It is true that, by the Ostwald-Freundlich equation, a larger crystal is thermodynamically favored over a smallerone; but the energydifferences in the usual particle rangesissmall and the activation energy high, so that the likelihood is rather low. If sufficient moisture is present so that the vapor pressure in the container exceeds that of a saturated solution, then some of the drug willdissolve in sorbed moisture. Fluctuations in temperature are never absent and would cause dissolution followed by precipitation, and this can lead to crystal growth. In cases wherea drug substance is capable of forming a hydrate, and where an anhydrate is used, growth by way of hydrate formation is possible. Ease of reconstitution is usuallycarried out subjectively, inthat atester carries out the reconstitution in the prescribed manner and records the length of time required to finish the operation. For this purpose it is important to have detailed directions on how the reconstitution is to be carried out, and to be sure that there is no operator-to-operator performance bias. To insure the latter, aset of operators is usually selectedfor the operation at a point in the stability history. These operators will then be the test instruments for all testing of reconstitutability of oral powders. The manner of screening operators could be as follows. A random sample is taken of a batch of a product. Random sets of four are taken from this random sample, and e.g. three operators tested. They are each given four samples to reconstitute on the first day, four on the second day, and four on the third day. It is a good policyto have two batches and mix them by day and operator, so as to carry out the test in a blind fashion. The results of such a screening could be as shown in Table 3. e 3 Screening of Operators for Reconstitution Testing. Reconstitution Time (min) Operator

Day 1 Day 2 Day 3

1.3 rt 0.3 1.7f0.5 1.5 f 0.6

1.5 k0.2 1.4f0.4 1.3 f 0.5

1.4 f 0.5 1.5 k 0.3 1.7 rt 0.4

12 10

Change of reconstitution time of a powder on storage.

ranges showndenote standard errorsof the mean. to show a significantdifferencebetween opera reconstitution time could change as shown in Fig. 16. As mentioned, the most common reason for increases in reconstitution time upon product storage is that the powder becomes more “lumpy” through cohesion developing over time or because it becomes coarser due to crystal growth. phenomena are associated with moisture content, and just as it is important to test the effect of the level of moisture content in the case of stability of a solid dosage form, so is it important to test it in the case of a powder. If m is the n u m ~ e rof mL of water adsorb~don one gram of powder, and if S is the solubility (inmg/mL) of all the soluble substances in the preparatio~,then, since the moisture layer isstagnant, the concentration of solubles at time t will be

c =(10.22) S[ 1 -exp(-qt)] where q is the dissolution constant (kA/V). he layer will have a higher viscosity(v), the more solid is dissolved,res sumably by a power function: q = p c

(10.23)

In analogy with the definition of viscosity, the force ( F ) needed to move two planes separated by a liquid isproportional to the viscosity. It would also be proportional to the amount of liquid, m, in the powder situation stated, so that combining this concept with Eqs. (10.22) and (10.23) gives F =m

S” [l-exp(-qt)]” (10.24) *

so the reconstitution time would be proportional to this. this pattern.

The physical properties associated with tablets are disintegration, dissolution, hardness, appearance, and associated properties (including slurry pH). For special tablet products (e.g., chewable tablets) organoleptic properties become important. These have been described earlier, but in the case of tablets, the chewability and mouth feel also become of importance. The properties will be discussed individually below. 8

he “hardness” properties of a tablet are usually assessedby subjecting the tablets to a diametral failure test. The tablet is placed (Fig. 1’7) between two anvils, one of which is stationary. The other anvil is moved at constant speed against the tablet, and the force (as a function of time) isrecorded. The force, at which the tablet breaks is denoted the “hardness” and is usually measured in kp (kilopond= kilogram

“7 (a) Hardness tester and (b)friabilator.

force). Other older units (Strong Cobb Units, SCU, or pound force) are used, usually when older instrumentation is used. Until recently, one limitation was that forces over 20 kp would simply register as F >20 kp. Newer instrumentation allows for quantitation of higher forces.From a stability point of view this is important, since the better a parameter can be quantitated, the clearer the picture that emerges will be. Tablets are made either bywet processing(wet granulation) or by dry processing (direct compression or slugging/roller compaction). In the former case a binder in solution is added to the powder mixture (or is contained in the powder mixture, and wetting then carried out). The binder forms soft bridgesbetween particles, and when the granulation is dried then these bridges become hard. They form the bonds during the compression, and this is one of the reasons for the addition of the binder. The hardness of the tablet is tiedin with the strength of the bond. The nature of the actual bond formation will be discussed presently. In order for a bond to form, the particles or binder bridgesmust first be exposed to stresses (pressures) that exceed the elastic limit of the material. On failure, the material will either deform plasticallyor experience brittle fracture. A material that flows well and has a low elastic limit is, therefore, easy to transform into a tablet, and several such materials, known as direct compression ingredients, are used in the manufacture of pharmaceutical tablets. In these casesdrug is simply mixed with the direct compression excipient(and other excipients), lubricated, and compressed. If the drug content is less than (approximately) 20% then the tablet will (generally) have the properties of the direct compression ingredient. At higher percentages, direct compression is usually only feasible the if drug substance itself is fairly compressible (i.e., has a low elastic limit). The hardness of a tablet will be a function of the strength of the bond and the number of bonds. However, this is statistically oversimpli~ed.If there are, for instance, many bonds in the bottom of a tablet and only a few in the top, then the tablet will break easily. Hence it is the average bond density and the standard deviation of the bonds that are really of importance. The same is true about the strength of the bond. Train (1957) has shown that the particle density in a tablet varies from spot to spot, and hence there is a variation in the density of the bonds (and probably in their strength as well). If the hardness of a tablet is plotted versus the applied pressure, then a plot such as shown in Fig. 18 results. It is seen that the curve goes through a maximum. For good formulations, this maximum doesnot occur until very high pressures(outside the range of pressures used in pharmaceutical tableting). The maximum occurs becauseabove the critical pressure, P*, the tablet will laminate or cap, and a laminated tablet (Fig. 19) willcontain strata of air and hence bethicker and weaker. Tablet thicknesses will respond in a manner opposite to the hardness, i.e., show a minimum (e.g., at 500 MPa in Fig. 18). The reason for this phenomenon isthe following: A s applied pressure increases, the number of bonds, N, increases as well. ut a s s u ~ i n gthat there is a m a ~ i ~ u ~ number of bonds, N*, that can be formed, then the strength, H, of the tablet will asymptote as well. In a simplified manner the relations would be N = N . [l -exp(-q.P)]

(10.25)

c 6 U)

a> Lf

200

400

600

800

1000

1200

Pressure (MPa) Tablet hardness versus appliedtableting pressure. (Graph constructed from data by Carstensen et al., 1986.)

Lamination

Tablet

Capped

Laminated and capped tablets.

If the hardness is assumed proportional to the number of bonds, then H = PlV = H [l

-(10.26) exp(-qP)]

where H is the capability of the tablet to withstand stress and p is a proportionality constant (Fig. 20). During the tableting process, when the upper punch is released, a stress is exerted on the tablet, and this stress (S) is the larger, the larger the applied pressure, S =.f (P)

(10.27)

At a given point, S becomes larger than H, and then fracture (lamination) occurs within the die, before the tablet is ejected. There is a second type of stress that occurs during compression, and this happens upon ejection from the die. Here, many tablets expand, and this expansion is a stress that may also exceed H, i.e., laminated or capped tablets are formed.

Time (days) Hardness as afunction of time in pharmaceutical tablets. (Figure constructed from data publishedby Ghowhan, 1979.)

On storage, this expansion can continue (Gucluyildiz et al., 1977), i.e.,a tablet may become softer on standing for simple expansion reasons. Expansion is rarely checked as part of a stability program, and the cited article is one of the few published attempts to measure porosity as a function of time. ~requentlytablets will become either softer or harder within short periods of time after manufacture. Figure 20 shows hardness as a function of time for a series of tablets reported by Chowhan (1979). side from the quoted instance of porosity changes and expansion, there are caseswhere crystallization of a soluble compound has occurred via the sorbed amounts of moisture in the tablet, This happens most often with very soluble compounds, and in such casesit is important to ascertain storage in a dry environment. A test that is now a requirement in the ICH Guidelines is storage in the final container at 40°C, 75% RH. During this test moisture is usuallyadsorbed by the tablets, and this can then cause softening of the binder bridge because of moisture uptake. At times, redrying will reinstitute the original hardness, Sometimes hardening occurs when the sorbed moisture causes recrystallization of a compound or excipient.

Softening can be associated withchemical interaction. Several furoic acids (Carstensen and Kotha~i,1983), whentableted with microcrystalline cellulose, will cause a specific interaction leading to the formation of carbon monoxide (rather than decarboxylation of the acid). Thisinteraction is not slow at 55"C, and it causes the tablets to crumble. At room temperature the effectisless pronounced yet significant. ince a tablet, when produced, is not in equilibrium, there willbe a redistribution of moisture. This could make the bonds of a lower or a higher moisture content, and there may for this reason be a change in hardness during a fairly short period of time after manufacture.

Moisture Content of Selected Excipients Water content Excipient 10.8 Sta-RX 1500 Solca-Floc 3.7CMC Celutab Microcryst. 2.9 Cellulose Polyvinyl2.9 pyrollidone

TGA.

CalciumKarl carbide

5.8

4.6

9.2 8.6

Fischer 10.4 6.4 14.9 9.0

3.4

4.7

5.4

6.4

Source: Table constructed from data published by Schepky (1974).

The moisture content of granules, when theyare made initially, is a function of their particle size. itkin and Carstensen (1973) have shown that when granules are dried, each is associated with one given drying time, t*. Since the drying (if it is countercurrent, or fluid bed) is a diffusional process, conventional diffusion theory predicts that the amount of moisture left in a granule, m, in relation to the initial amount, mo, is given to a first approximation by (1 0.28) where LI is the diffusion coefficientof water in the granule and a is its diameter. The larger granules will hence, have a higher moisture content at the beginning, but the moisture will equilibrate, in most cases, on storage. However, oglio et al. (19’75) have shown that in some cases (spray dried sucrose granules) there willbe no redistribution of moisture between larger and smaller granules. The moisture contents of various excipients have been reported by Scheplcy (19’74) and are listed in Table 4. In stability situatio~sit is often the change in moisture as a function of storage time that is of importance. In such cases (Shepky, 1974), the thermo~ra~imetric method may be of advantage. It is of interest, in cases where moisture equilibrates and causes change in hardness on storage, to be able to assess the extent of moisture transfer within the tablet. A s mentioned at an earlier point, the situation is that (in a simple case of a two-component tablet) the two components (I and 11) have different moisture isotherms. These are approximated in linear fashion in Fig. 21 This is bestillustrated by example. ~ ~ a m10.3. ~ l e

tablet is made of two components, I and I .It has a given moisture content, 10 mg of water per g of dry tablet weight. The moisture isotherms are asshown inFig. 21. Calculate the final moisture contents of the two components after the moisture has equilibrated.

/ y = 0 . 2 5 ~ R = l .OO Component 2

2 3 a ._. C . ’

P 0

Component 1 20

Percent Relative ~ u ~ i d i t y

Moisture exchange between ingredients.

Answer.

For component 11, 1Omg of moisture per g of dry solid corresponds to a relative humidity of 40%, i.e., the equation for the isotherm for 11 is given by (10.29) since it passes through the point (10, 40). The moisture isotherm for 1 has the equation 300 y1 = 1 5 x 1 = 20x1

(10.30)

since it passes through the point (15, 300). At the tablet moisture content (lOmg/g dry solid) it is in equilibrium with a gas phase of relative humidity 10 x1 = = 0.5% RH 20

(10.3 1)

Hence the situation is not an equilibrium situation, because there is no c o m o n vapor pressure over the solids. Component I will, therefore, give up ( q grams of)water, and 11 will pick up (q grams of)water until a common vapor pressure (P) in the porous and external vapor space has been achieved. The equilibrium relative humidity is given by

X*= 0.25(10 -4)

(10.32)

X* = 20(0.5 + 4)

(10.33)

and

and equating the two right hand sides then gives

2.5 -0.25q = 10 20q

(1 0.34)

20.25q = 12.5

(10.35)

or "4'

12.5 20.25 -0'6mg

(10.36)

It simplifies the computation that the moisture content is givenin mg of moisture per g of dry solid (i.e.,not in percent, which would berelated to mg of moisture per g of total weight). It should be noted that this situation is simplified by assuming the isotherm to be linear.

Tablets (whether coated or not) are usually subjected to a disintegration test. The disintegration was the first in-vitro test used by the U.S.P. It is now not obligatory compendially (but is recommended); in an obligatory sense it has been replaced by the dissolution test. This latter, hence, is the more important test, but it will be seenthat there often is a correlation between the two, and since the disintegration test is much more easily carried out, a stability program will check disintegration frequently, and dissolution less frequently, primarily due to labor intensity. in Fig. 22. It is The apparatus used (U.S.P. XX, p. 958) is shown schematically an apparatus where sixtubes are placed in holders on a circular screen, which is then raised and loweredbetween 29 and 32 times per minute through a distance of 5.3-5.7 cm in a 1000mL beaker containing the disintegration medium (either water or N/ 10 hydrochloric acid). The wire mesh oscillates so that it is 2.5cm (or more) below the surface at the upstroke and 2.5cm (or more) from the bottom of the l

Oscillating Shaft

Fig. 22 Disintegration apparatus.

1000mL beaker at the downstroke. The open-ended glass tubes are 17.75~0.25 cm long and have an inside diameter of 21.5 cm. The glass thickness is 0.2 cm. Each tube is provided witha disk 95 mmthick and 20.7 mmin diameter, made of plastic of a specific gravity between 1.18 and 1.20 g/cm3. here are five 2 m holes in the cylinder(one of them in the axis). The disk a1 has notches in it and serves to keep the tablet within the tube and submerged during the stroke of the assembly. To operate the apparatus,one tablet is placedin each of the six tubes, disks are added, and the apparatus is operated at 37°C in the immersion fluid. For quality control release purposes as well as for investigational purposes the time is noted when all tablets have disintegrated completely, and if not all tablets have disinte~rated at the end of the specification limit, then the basket isremoved and the tablets observed, If one or two tablets have failed, are tested, and these must all disintegrate within the limit. it is important to note the time that each individual tablet disintegrates. should be pointed out thatcomplete disintegration is definedas “thatstate in ny residue ofthe unit, except fragments of insoluble coating or capsule shell, remaining on the screen of the test apparatus is a soft mass having no palpably firm core.” There are apparatuses on the market that have a sensor attached to the disk and can determine this state automatically and record the time at which it occurred. uch an attachment is strongly recommended for stability studies, since it provides n easy means of recording the time of disintegration of each tablet. There are relatively few articles in the pharmaceutical liter~turethat deal with the subject of the change in disintegration and dissolution upon storage, yet these qualities are as important as the retention of potency of the active compound. If roduct falls short of specifications during its shelf life,it becomes unsatisfactory, of the particular parameter that is shortfall in^. fairly systematic study of this is the work by Chowhan (1979) and dissolution times of e.g. dicalcium phosphate based ta studied for prolonged times at 25 and 37°C. The pattern is a sigma minus type of pattern as shown in Fig. 23. Carstensen et al. (1980a, 198013) have shown that there often is a correlation between dissolution and disintegration, and Carstensen et al. (1978a, 19788b) have shown the theoretical basis for this. igure 24 shows sucha correlation of dissolution and isi integration times in a U.S. Couvreur (1975) has shown that the disintegration of a several factors. If the tablet disintegrates by virtue of a disint once it is wetted, then the most important attribute is the rating liquid penetrates the tablet, and hence the contact angle between the solid and the liquid is of importance.

Cruaud et a1 (1980) showed that there was a direct correlation between dissolution and porosity in a case where the correlation between disintegration and dissolutio~ was not apparent. There is one welldocumented case ( ~ u c l u y i l det i ~ al., 1977) where the porosity was shown to change in a tablet as a function of time. hat this indicates is that if

L-

0 0

Minutes

mg Dissolved 3 Top figure: Correlation betweendissolution and disinteg tion upon storage. constructed from data published by Carstensen et al., 1980a.) ttom figure: Direct correlation between dissolution and disintegration.

dissolution and disintegration change on storage, then they may befunctions of the change in porosity, if indeed porosity changes as a function of storage time. A rational way of studying this would be to study mercury intrusion as a function of time intablets of a drug, and to study simultaneously the disintegration and dissolution profiles. There is a distinct effect of moisture uptake or equilibration on disintegration (and hence, indirectly, on dissolution). If the liquid penetrates an “average” pore, then it encounters, on its way, disintegrant particles. It is assumed that there are q disintegrant particles per linear length of pore. It is also assumed that N particles (per pore) must have been wetted before the tablet can break up (disintegrate),

70 .W

E..

50 40

PercentRelative

100

Hu~i~ity

Effect of storage at various relative humidities on disintegration time of an antidiabetic tablet. (Graph constructed from data published by G r i m and Schepky, 1980a.)

According to the Washburn equation (Washburn, 1921; Nogami et al., 1966; Carstensen, 1980), the length, L, o f penetration of liquid at time t is given by

L2=(r *f2rl cos (b) t = P . d t a

(10.37)

where f.cos(b*r

(10.38)

and where f is the interfacial tension, d is the average pore diameter, r is the pore radius, (b is the contact angle, and y is the viscosity. The number of particles wetted n, is related to L by q=-

n L

(10.39)

n L=4

(10.40)

where q is a proportionality constant. The disintegration time, tN, occurs where n = N , SO

N LN=4

(10.41)

and (10.42)

~ysicalTestin

Combining these two equations gives (10.43) Hence, the following hold for the disintegration time,

tN:

1. It is the larger the more disintegrating particles must swell to make the tablet disintegrate. 2. It is the longer the finer the pore (the smaller d is). 3. It is the smaller the larger the disintegrant concentration, q. 4. It is the smaller the larger the value of P(the smaller the contact angle and interfacial tension). Of these, N may change, e.g., if the disintegrant becomes wetter, and partly expanded as a result of moisture uptake, this will affectthe disintegration adversely. For instance, the Joel Davis test (40"C, 75% RH for three months) has an adverse effect on disintegration for this reason, although it isonly true if the relative humidity of the testing station is above a certain critical moisture content (Grimm and Shepky, 1980a). This is demonstrated in Fig. 24.

The dissolution apparatuses used are usually USP Method I (basket apparatus) or USP Method I1 (paddle apparatus). Carstensen et al. (1976a,b) have pointed out that the hydrodynamics of the basket method is poor and results in highly different liquid velocities in different parts of the apparatus, and also causes a phenomenon known as coning: powder accumulatesat the bottom of the dissolution vessel, where it is fairly stagnant and hence dissolves slowly. Most tests nowadays are therefore carried out with the paddle apparatus. The assembly is describedin USP XX p. 959 and is basically as shown inFig. 25. The original apparatus could be operated at SO, 100, or 150 RPM, but the more up-to-date apparatus has a variable speed rheostat. In almost allinstances the

900 rnl..Dissolution

Paddle (Agitator) ablet Fig. 25 USP dissolution apparatus.

A asks for 50 RPM (sometimes100 RPM), but onlyrarelydoes it accept 150 RPM and insist on the test being “discriminatingp’. The apparatus should be subjected to a suitability test (USP XX, p. 959), using one tablet of the USP dissolution calibrator, disintegrating type, and one tablet of the USP dissolution calibrator, nondisintegrating type. The apparatus is satisfactory if the data are within the stated range of acceptability for each calibrator. The dissolution medium is water, hydrochloric acid, or pH 7 buffer. These should be deaerated, since dissolved air may interfere with the dissolution rates. The procedure used isto transfer usually 900 mL of the dissolution medium to the dissolution vessel and bring it to 37°C. After temperature is equilibrated, the thermometer is removed and one dosage unit is placed in the apparatus. Care is taken to exclude air bubbles from the surface of the tablet and to operate the apparatus right away. After given times, samples are removed from the supernatant and assayed, and the concentration is plotted as function of time. The results may be expressed as percentage of the highest possible concentration (D/V, where D is the dose and Vis the volume of the liquid). Monographs specify, usually, a given time at which a certain percentage of label claim, Q, must be dissolved, and the term Q 3 0 for instance is frequently used; it indicates the percent of label claim dissolved after 30 minutes. This is known as a one-point assay. For quality control purposes, this is acceptable, but for stability purposes, if rational graphing is contemplated, a dissolution curve rather than a one-point determination should be determined, This will allow determination of the dissolution constants, which will be discussed shortly. It is often (particularly with slowly dissolvingor sustained release products) of importance to have the value “at infinite time.” This is usuallyimitated by increasing the rotational speed (e.g.,to 150 RPM) and running the dissolution for an extra two hours, It is, in this scheme,assumed that all the drug willdissolveundersuch circumstances. Shortcomings of the apparatus are still (a) that tablets made with excipientsof high density will havea tendency to “cone,” i.e., after disintegration accumulate in the dead spot just below the agitator; this gives false lows in dissolution rates, and (b) that capsules (and some tablet formula) may float. To avoid floating, a coil is usually placed about the capsule. It is interesting that some tablet formula with relatively small changes incomposition (or compression pressure) will changetheir density, so that they float in one omp position (or pressure) and sink in another (only slightly different) composition (or pressure). Expansion of tablets during storage may also change the density so that a table can change from a sinking to a floating composition which will give rise to anapparent slowing down ofthe dissolution rate. It should be pointed out that dissolution testing of pharmaceutical products is carried out for several different reasons. In the early stages,the intent of dissolution testing isto get a feel for the comparative estimated bioavailability (on a rank order scale) of different formulations. In preformulation, intrinsic dissolution rate constants are usually estimated. Although it is not possible, in a direct manner, to tie this in with an estimated bioavailability, it gives a feel for whether the drug substance will be exceedingly problematic, very problematic, problematic, or (in rare cases) not problematic. This feel is comparative with the intrinsic dissolution properties (obtained in a similar fashion) for other drug substances previously developed.

Relative Rankings of Furosemide In-Vivo VersusIn-Vitro (Random Cross-Over, 12 Patients)

1 2 3 4 5

D A=C A=C E

*Smallest 1” gets best rank, highest AUC and C gets best rank (i.e., lowest number in column 1). Source: Table constructed from data published by McNamara et al. (1987).

In formulation, it is generally assumedthat if e.g. three formulae, A, (Table 5) ar oped, then the one that has the fastest dissolution rate should be the best. er this general statement is correct is debatable (Table 5), but her criteria it is an accepted yardstick. ostformulation9i.e. at the point where the new product is manufactured, the new product has beco an established or old product, dissolution is in the domain of quality assurance. re it is part of a specification, and the intent of conducting the test is to declare to the public (given the criterion that in-vitro dissolution within certain limits corresponds to in-vivo performance)that the product made on day X, year U, is comparable to (and should perform in a manner similar to) the batches made year previously on day 2,year Q, when it was tested in the clinic. If this premise were generally correct, then an in-vitro dissolution test would be universal for all formulae, and it will be seen below that that is an ~nwarranted extension. Thequestion whether batches of the same formula fall under such a rank order rule isprobably acceptable. In the history of a product, however, small changes S shift the in-vitro to are often made, and the question whetherthesesmall in-vivo interrelation is, of course, not known a constitutes smallness is not clear (and actually is not determinable). es are defined, now9 as changes that tighten specifications and do not involvechangein procedure, t, or raw material. a “substantial” (major) formula change is made, then the bioequivalence clinical formula and the new formula must be established. equent formula changes are made (e.g., a bioequivalence study is carried r U, then another in year U 1, etc.), then comparisons should be made with the original clinical formula, not with the previous formula. If the formulae f successive comparisons were made are denoted A (clinical), B, C, etc., could be 0.8 times PA (and deemed and performance were denoted P, th deemed equivalent), but quivalent. Hence, equivalence testing the original clinical batches that were of the medical scheme(and the results of which wereapproved by the g Administration).

~rstensen

When a solid is compressed, then one might imagine that at“full” compression, the tablet would be similar to a perfect crystal, in that there would be no void space left in it. This is never achieved, however, and the fraction of void is called the porosity. This may be visualizedas isolated pockets of void spaceor, asthe porosity increases, strings of void, eventually terminating at the surface. The porosity at which this latter situation is achieved is denoted the threshold value. Threshold values for a drug and its excipients in combination are important because they govern such properties as dissolution, hardness, and disintegration. For this purpose, percolation studies are often employed in pharmaceutical research. Leuenberger and Leu (1992) and Leu and Leuenberger (1993) introduced the concept of drug percolation to the pharmaceutical sciences. By this, a pharmaceutical system is described as a bondIsite system. In this concept, a cluster is defined as a group of nearest neighborsiteswhereall positions consist of the same component. There is a concentration where there is maximum probability that the clusters will start to percolate, and this is the percolation threshold. If the measured porosity of the tablet is denoted and (after dissolution) the porosity created by loss of dissolved matter is denoted E*, then the so-called p property is

p = ”CEc

0.44)

E* is the initial developed (matrix) porosity, c is a constant, and cc is where E = the critical porosity threshold for percolation. This tiesin with the Wiguchi type plot, the slope of which is b, and p is defined as

p = [2A -E q ” 2

(10.45)

where A is the drug load (g/cm3 of total tablet) and S is solubility. When porosity is plotted versus /l value, then a straight line ensues that cuts the x-axis at the percolation porosity. The threshold for drug percolation may be obtained when more drug is available than that described in Chapter 9. Soriano et al. (1998) have described percolation methods that are done primarily by conducting dissolution studies with drug substance at various concentrations. Theyemploy the method ofBonny and Leuenberger (1993) and Leuenberger and Leu (1 992) for this purpose.

In post NDA testing, there is somereason for not carrying out dissolution at more than one timepoint, because of both human resources and equipment. In pre-NDA situations, however, as described e.g. by Prandit et al. (1994), the importance of carrying out multiple time points in dissolution cannot be stressed enough. Conclusions are difficult to reach if this is not done. For instance Prandit et al. (1994) reported that aging affectedthe dissolution of nalidixic acidtablets and concluded that the effect wasnot attributable to anincrease in disintegration time (as measured in a dissolution apparatus). Published data often

hysical Testin

suffer from being one point data, so dissolution/disintegration correlations cannot be deduced from the reported figures.

There is alwaysthe problem of what dissolution medium to use. For poorly soluble drugs there are several approaches: cosolvents, micellar systems, and/or large dissolution volumes. Naylor et al. (1993) studied the mechanism of dissolution of hydrocortisone in simple and mixed micelle systems by using a rotating disk and found the Levich equation to hold.

The problem of whether an in-vitro dissolution test generally measures in-vivo performance on a rank scale basis is still open to debate, when the problem is considered in general, i.e., if product A from manufacturer A has a dissolution rate curve “above” that of manufacturer B, willhis product also have a better in-vivo performance as far as large magnitude (Cmax) and short peak time(Tmax) for the maximum of the blood level curveand high valuefor the area under the blood level curve(AUC)? The general premise isthat the answer is yes,but as shall be seen below, this is not necessarily so. The correct general statement is that if two batches of the same product and formula are tested, then such a comparison is correct, i.e., that a “higher” dissolution curve implies at least one of the following: lower Tmax, higher Cmax, or higher AUC. An example of noncorrelation, when the formula is not the same, is the work by McNamara et al. (1987), in which furosemidefrom five manufacturers was tested against a solution. The relative rankings are shown in Table 5. It is seen, then, that the best performer in vivo (I)) is by no means the best performer in vitro, and that the worst performer in vitro (E) is not the worst performer in vivo. The best and simplest method for correlation of in-vitro to in-vivo data would appear to be the mean residence time (MRT), and such comparisons have recently been described by Block and Banakar (1988). MRT is defined by many authors as shown in Fig. 26. The MRT factual definition isa measure of the “average” length of time a drug molecule is in the body (Fig. 26). Mean residencetimevia statistical moment has also beendescribed by Yamaoka et al. (1978). Podzeck (1993) has compared in-vitro dissolution profiles by calculating mean dissolution time and mean residence time. Of late, deconvolution has been often reported and may form part of the 1995 USP. This method consists of comparing a blood level curveafter solid dosageform administration with one after either solution or IV administration. The amount dissolved in the GI tract is then obtained by deconvolution. Sugawara et al. (1994) tested a series of controlled release preparations of prednisolone in alginate gel beads, all in a drug-to-alginate ratio of 1 :4. As seen in Fig. 27, they were able to obtain in-vitro methods that “matched” the amount released in-vivo. It is seen in the figure that for the fast releasing formulation (a), the in-vitro test, whether at pH 1.2 or at pH 6.8, follows the deconvoluted in-vivoresults fairly well, but for the slow formula, it is only the pH 1.2 in-vitro test that correlates with the deconvoluted in-vivo dissolution test.

100

80 60

40 20

Hours Cumulative blood level curve (or urinary excretion curve, or dissolution curve).

In-Vivo Czeconvoluted

4 6 Time (h)

pH 12.2 and 6.8

7 Time vs. percent released.(Graph constructed from data published by Sugawara et al., 1994.)

The deconvolution method used wasthe one that has been describedby Katori et al. (1991). Other, more recently developed methods are those of Gillespie and Cheng (1 993). They firstcreated a hypothetical clean curve withconvolution. Then the absorption rates and cumulative amounts absorbed of the drug and metabolite were estimated by the proposed deconvolution method. For this purpose, polye~ponentialfunctions werefitted to the simulated data. The resulting parameters were compared by a multidimensional deconvolution program NDC (user-frien~lyIB

The problem from a stability point of view is that at times the dissolution curve will change as a function of storage time [as e.g. shownby Chafetz (1984) for hard shell capsules], but the bioavailability “stays the same.” In such a case the in-vitro

T (or dissolution curve)wouldchange but the in-vivo(either deconvolution curve) wouldnot, so how can there be a correlation between the two? The accelerated test inthe ICH Guideline (40°C,75% RH) is too severe a test for hard- and soft-shell capsules. Upon dissolution, a skin (a pellicule, or as some authors call it, a pellicle) will for round the capsule in the dissolution apparatus, and this willprevent dissolutio r instance, Dey et al. (1993)exposed etodolac capsules to the accelerated test so that they formed pellicules and showed that the dissolution was not affected whentested with enzymes,but that pellicules formed and dissolution decreased drastically on storage when nonenzymatic fluids were used. They showed that there was no difference in blood level curves of fresh, stored, and failed batches. In the case of hard- and soft-shell capsules,gelatin can interact with substances in the fill. Cautum and Schott (1994) demonstrated an interaction of anionic compounds (substituted benzoic and sulfonic acid dyes) with gelatin. Capsule fills that contain, or on storage produce, keto groups will always show this phenomenon (Carstensen and It should be pointed out that when disintegration of a dosage form changes on storage, it usually happens quite rapidly (usuallywithin12weeks) at room temperature. Often, however, the tablet is not checked until 6 months after manufacture. There are then instances whereit would seem to be logical to attempt an accelerated test at higher temperature. There has, to date, not been any convincing correlation between disintegration (and dissolution) profiles at higher temperature, vis-a-vis those at lower temperatures. Judging from the factors that affect these two properties, this is not surprising. But what is more to the point is that changes can usually be determined rapidly at room temperature. It is therefore more rational to determine disintegration at 4,8,and 12 weeksat room temperature in stability programs, and to dispense with testing at higher temperatures. Cordon et al. (1 993)have reported on the effect of aging on the dissolution of wet granulated tablets containing superdisintegrants. Often the decay indissolution efficiency is due to the lengthening of the disintegration time. There is, obviously, a correlation between particle size and dissolution, and if the particle size changes as a function of storage time, there may be a correlation between accelerated temperature storage and dissolution. But in such a case the correlation should be established on the neat drug, as was done e.g. by Grimm and Shepky (1980b). Their data for oxytetracycline are shown in Fig. 28. Dukes (1984) and Murthy and Ghebre Sellassie, (1993) have discussed storage stability of dissolution profiles in general, and Rubino et al. (1985) have described the specific storage stability of the dissolution of phenytoin sodium capsules.Carstensen et al. (1992) have discussedthe mathematical basis for change in dissolution curves of dosage forms as a function of storage time. They employthe sigma minus model for dissolution, i.e., M

= l -exp[--k(t -ti)] MO

(1 0.46)

where t is dissolution time, ti is dissolution lag time, MOis initial amount in the dosage form, M is the amount left undissolvedat time t, and k is a dissolution constant (time"). ti is primarily a function of disintegration,time*

cars tense^

Particle growth in acceleratedstorage of oxytetracycline. (Graph constructed from data published by Grimm and Scbepky, 1980b.)

40 Decreased k 120

4 02 0

60 80 Time(min)

100

120

Dissolution upon storage.

Frequently, upon storage, t i may change, but k may not, in which case the dissolution curves simply move in parallel to higher and higher mean dissolution times. k , however, may change, and t i may stay constant, in which case the curve becomes "flatter" (Fig. 29). Finally, both may change, giving rise to a flattening and a parallel displacement of the curves. If such parameters as t90 (the length of time for 90% to be dissolved) or Q45 (the amount dissolved at 45 minutes) are employed, then power function relationships result, and these are difficult to interpret. A better approach is to study and plot k and t i as a function of storage time. If t i on storage approaches 45, then the storage stability curve may have a shape as in Fig. 30.

311

hysical Testin

y = 3.5291

- 9.3878e-2x

RA2

= 0.994.

Storagetime(months) Fig. 30 Storage stability of lag time. For instance, if a secondary parameter, such as the time needed for 50% dissolution, t50, is followed, then (10.47) -ti) In 0.5 = -0.693 = "k(t50 or (1 0.48) and if

where ty is initial lag time,ty is lag time at infinite storage time, cf> is storage time, and q is the stability constant, then (10.50) If, now, the storage stability of k is of importance, the expression for t50 becomes even more complicated (but can obviously be deduced). Jmgensen and Christensen (1996, 1997) have approached this problem by introducing a so-called Order Model. By this an order of reaction, n, is assigned to the dissolution curve, and the expression becomes M

== 1 -[ 1 -{(1 -n)k(t -f(to)))]"'""' (10.51) MO

where f ( t 0 ) is the lag time function given by f(to>

= t o [ l -e x p ( ~ ) ]

(10.52)

t is, again, the dissolution time.

A stability program should record the appearance of tablets as a function of storage time. This is most often done by subjective description, or by a rating index (0 for unchanged, 5for vastly changed).Quantitative methods exist and are the following: Comparison with color chips or charts (Rothgang, 1974) lving the dosage form and measuring the solution spectrophotometrically mouda and Salakawy, (1971) Photography (Armstrong and Marsh, 1974) eflection measurements(Matthews et al. (1974/75), Carstensen et al. (1964), Carstensen (1964), Goodhart et al. (1967), Turi et al. (1972), Wortz, R, In the case of the second and fourth methods, a qualitative appearance description is always necessary, becausethe instrument will “average” the product. Comparison with chips can be used but issomewhatsubjective.Such color charts have triangularly arranged chips, and the operator matches the object with a chip, which has a coordinate number. In fact the degree of whiteness (L), redness, (a) and yellowness (b) can be calculated from this, and it will be seen later on that this will allowfor quantitative treatment of the change of the color of a pharmaceutical tablet or capsule. Photography, of course, isrelying on stringent adherence to conditions (exposure, aperture, and development) to insure that it is actually the tablets that are being compared, not the procedure for making the photograph. Reflection measurements are often carried out in tristimulus meters and have been used quite extensively with varying degrees of success. If a tablet (or other surface) is placed in the meter, then reflectance values at three spectral regions are registered and recorded as x,y and z values. Rowe (1985) has reviewed these -3 Y, and the yellowness index and points out that the whiteness index is 4( lOOZ/Zo) is 100[1- (lOOZ/{ Y ZO}), where ZO= 118.1. In actuality, the degree of whiteness, L, the degree of redness, a, and the degree of yellowness, b, are given bythe formula (for unter tristimulus meter):

= 175

b =70

(X/98.041) (Y/lOO) (Y/100)”2

(Y/lOO) -(Z/118.103) (Y/100)1/2

Tristimulus Parameters for Some Colors X

Color White 65.0 yellow Light 32.9 ochre Yellow Scarlet Magenta Turquoise Emerald green

101.1 7.0 7.5 4.3 23.9 23.9 13.8 17.7 12.3

84.3 69.9 28.7 20.6 13.0 21.o 30.1

44.7

91.8 83.5 53.6 45.4 36.0 45.8 54.8

-0.7 -7.4 16.l 62.2 55.2 -26.5 -38.3

-1.0 53.5 29.3 26.2 -13.9 -25.7 25. l

Days at 70°C

31 Reflectance X values (with initially X= P)of vitamin C tablets as a function of storage time. (Graph constructed from data published by Carstensen et al., 1964.) Kornerup and Wanscher (1967) and Rowe (1983) giveas an example the following values for some standard colors (Table 6). The values calculated in the last three columns by Roe correspond quite well with those obtained from or listed in color charts (Komerup and Wanscher, 1967)). It should be pointed out first of all that reproducibility in reflectance meters is poor, and so results should always be obtained as averages of at least nine independent measurements. Since these are rapidly carried out, the labor is not all that intensive. Changes in these values are difficult to interpret from a qualitative point of view, but the following procedure allows extrapolation, using x, y, or z (or composites). Carstensen et al. (1964) have shownthat the response values(Y) can be plotted as a function of storage time ( t ) to give graphs as shown in Fig. 31. This type of plot can be plotted as a sigma minus function: Y' =(10.53) YW{1-exp(-kt)}

rstensen

314

The k values can be plotted as an Arrhenius plot, i.e., one may, after short periods of time, at elevated temperature, calculate an extrapolated k value at room temperature. By sampling daily at 55”C, one can determine the Y value (%owe,limit), which corresponds to the poorest appearance that is acceptable. Since k is known for room temperature (k25),it is possible to calculate a “shelf life date” (t*) based on appearance from inserting Eowerlimit into Eq. (10.54): (10.54)

There are several types of sustained release principles used inpharmaceutical products, and a detailed description is beyond the scope of this book. What will be done here is simply to state the types of dissolution profiles that can be expected, and how the parameters could change with time.

The coated nonpareil seed isthe original sustained release form invented by SKF in the 1950s. Here a drug is applied (in the form of a sugar syrup) to monodisperse sugar crystals. Drying is carried out after each application step, so that the drug eventually isin a sugar matrix around the original seed. This beadlet isthen coated with either a semipermeablefilm or an impermeablefilmwith a solublefiller. The latter, upon exposure to dissolution medium, will allowthe soluble fillerto disin through solve, so that pinholes are created in the film.Liquidthendiffuses the film (or the holes in it), becomes saturated on the inside of the beadlet, and the dissolved drug then diffuses out. Thediffusion takes place under an (approximately) constant concentration gradient (the solubility of the drug in the medium), as long as there is undissolved material inside the beadlet (and the concentration is lowin the outside fluid creating sink conditions). Once the last drug has dissolved, the concentration inside the beadlet will decrease, and the diffusion slows down. It is, therefore, often, difficult to get the last 5--10% of material to release from this type (and other types of sustained release) dosage forms. There are, obviously, three stages in the dissolution (Fig 32): 0

ti: Penetration of liquid into the pellet. ti is the time it takes for this to complete, and it is denoted the lag time. ti c t tf: tf is the point in time where all the drug inside the pellet has dissolved. t >tf: This is the finalperiodwhere dissolution isslower. t

The general dissolution pattern in the period ti < t

tf is (10.55)

M is the mass not dissolved (and MO is the dose) and is obtained by multiplying

concentration with dissolution liquid volume and subtracting this (the amount dissolved) from MO.k is the dissolution constant and willbe the smaller (and ti

hysical Testin

l00 80 60

40 20

0 0

.32

Hours

Release patterns of thinly and thickly coated pellets.

Time (hrs) Fig. 33 Dissolution curve of sustained release beads. The thickness of the coat of I1 is 1.6 times that of I. The graph shows the decrease in slope and the increase in lag time with increased thickness. (Graph constructed from data published by Carstensen, 19’73.)

the longer), the thicker the film and the lower the amount of soluble fillerin the film. Figures 32 and 33 show this type curve(Carstensen, 1973). The first twopoints have been omitted, as have the last three points, i.e., t i =about 1.5 hours and tf =about 5 hours. The points omitted are such that the remaining points give the best linearity. It is seen from the graph that the least squares fit equations are I

ln[Ml] = 5.09 -5.04t

(10.56)

h[lM2] = 4.8 -O.29t

(10.57)

so that, as expected, the slopes are in a ratio of 5/3, i.e., the inverse of the ratio of thicknesses. The actual lag times are found by setting ln[MI =ln[100], and they are tl= 0.67 hours and tII =0.96 hours, i.e., again in the correct ratio. tf is the point of inflection, i.e., occurs when all the drug inside the pellets will have dissolved (although not all will have diffused out). In stability programs, ti and k are the logical parameters to follow, i.e., complete dissolution curves should be determined. Again, it is wise to do this at room temperature storage at fairly short intervals at the onset (4,8, and 12 weeks). Again, accelerated testing is not of much use. let

Tablets can be made ofe.g. a waxy substance, whichdoes not dissolve or disintegrate, but erodes away. The drug in the eroded portion will dissolve, and (in theory) the drug in the noneroded part will not have dissolved.There is, however, always somepenetration of liquid into the waxy tablet, so that more than the eroded drug will often have dissolved. Ifpure erosion occurs, then the dissolution equation will be

where Ke is an erosion constant (cube root dissolution rate constant) and ti is the length of time of wetting. Both of these parameters can be calculated at different storage periods, and changes can be monitored in a logical fashion. Accelerated studies of this are not meaningful.

where (10.61) a is here the surface area through which the diffusion takes place, E is the porosity, and A is the loading, the amount of drug per cm3 of dosage form. E, the porosity, is the inherent porosity of the tablet plus the porosity created by the drug that has dissolved (i.e., A l p , where p is the density of the drug). Eq. (10.61) applies to situations where the drug dosage, A,is larger than SE^. If this is not the case (Table 7), then the equation takes the form (Fessi et al., 1982)

= a2Dt

(10.62)

le ' 7 DissolutionAccording to Eq.(10.62) Square Time (mi$

0 8 15 45 78 96 128 164 216 276

Amount

0 12.3 19.8 43.4 59.4 67 75.5 80 84.9 87.7

root of time (min1f2)

0 2.87 3.87 6.71 8.83 9.80 11.31 12.8 14.7 16.6

Amount released

0 151 392 1884 3528 4489 5700 6400 7209 769 1

Source: Data from Fessi et al. (1982).

Certain products are not porous but depend on the dissolution of the drug to create the porosity. In such cases there is a minimum drug content necessary for creating a porous network, and some of the drug will be occluded, i.e., will never release. A practical minimum is about 20% drug is such cases. The derivation of Eq. (10.59) is basedon the assumption that the penetration of liquid isfaster than the dissolution of the drug. If, e.g., the contact angle (wettability) changes with storage (e.g., due to moisture redistribution), then this assumption could be rendered false. Equation (10.59) applies only as long as there is undissolved material in the matrix (and until liquid has penetrated into the center of the tablet) The parameters Ki and ti may be monitored at various periods of room temperature storage time. In the case of insoluble matrices, accelerated studies might be possible in certain instances (i.e.,whenneither matrix nor drug changesphysically at the higher temperatures). Table 8 gives an example of Eq. (10.62). These data aredepicted graphicallyin Fig. 34. It is seenthat the least squares fit is given by

@(10.63) = -2.55

+ 0.477t

when the linear points are used. [Theseare, again, obtained by successively omitting terminal points (beginning and end) until the best linear fit is obtained.] It is noted that the first twoand the last three points have beenomitted, i.e., ti =2.55/0.477 = 5 minutes and tf (from the best, high point omitted) is 128 minutes. Again, 75.5% are released at this point, and this is quite characteristic, and it calculates out well for most such dosage forms as the point dissolved at the time the tablet has filled up with dissolution medium. Curing of the product is at times necessary. The work ofOmelczuk and cCinnity (1993) has, for instance, shown that matrix poly(DL-1actic acid) change release pattern if thermally cured.

60 0 0 F-

B 40

-z 20

100

200

300

Time (min)

Dissolution pattern in matrix dissolution. (Graph constructed from data published by Fessi et al., 1982.)

the tablets is neatly square root of time dependent as predicted by the Higuchi equation (Higuchi, 1962) shown in Eq. (10.58).

The osmotic pump is a tablet coated with an imperviousfilm, into whichis (laser-)drilled a hole of exacting dimensions. Dissolution liquid will penetrate into the interior of the tablet, and a saturated solution will form. The excipients are chosen so that they have a given solubility and hence produce a given osmotic pressure.(The drug itself contributes to this as well.)Thiswill be larger than the osmotic pressure in the outside liquid, and the difference between the osmotic pressure inside and outside will be the driving force by which liquid is being forced out through the hole. This givesrise to zero-order kinetics (which biopharmaceutically isan advantage), and the osmotic pump in many experimental situations, as well as in marketed situations, seems to be the dosage form that gives the most desirable release pattern, and also the one most likely to give in-vivo results that are predictable from in-vitro data. Dissolution data will therefore plot linearly, when amount released is plotted as a function of time. There may be a small nonzero (negative y-) intercept, i.e., a lag time. There will also be a point in time when there is no more solid drug inside the tablet, and deviations from linearity will occurfrom this point on. There is no literature published on the stability pattern of this type of dosage form, but it is to be expected that it would be no more (and probably less) prone to change on storage than the other types mentioned.

There are sustained release tablets that rely on gel-forming substances to accomplish the sustained release. In these cases the dissolution liquid will form a gel when it

encounters the surface of the tablet. rug must dissolve and diffuse out through this Astimegoes on, the gel layer gets thicker, a diffusion path becomes he data can be represented by Eq. (14.49baet al., 1979). There is also the possibility of some “sloughing ofP’ (i.e.,erosion) of gel, and in this case the release becomes a hybrid between erosion and diffusion through increasing thickness of gel.

ablets are often film coated and, less frequently, they are sugar coated. d, provides an excellent moisture and quite an coating does the same, but not quite as effectively. S are more stable in a (properly made) coated tablet, For instance, vita film coated tablet, of course provided there are no drastic incom~atibilitiesin the core. Film coating offers many advantages and often is the coat of preference, because (a) its application is much less labor intensive (cycle times being in hours for film coated tablets, in days for sugar coated tablets); (b) they also provide the advantage of allowing n engraving to “showthrough,’’i.e.,identification requires no extra operation. n the contrary, sugar coated tablets, for identification, require a separate printing step; and (c) there is an inherent advantage in film coatin in that it allows the appearance of a deep color without the use of much dye. Ifa uncoated tablet is colored, the dye is present throughout the tablet, whereas in a ated tablet it is only present in the outer layer (the film itself). nteric coated tablets belong in the category of coated tablets and will be treated below as well.

Film coated tablets are produced either in a coating pan or by column coating ost coatings, nowadays, are aqueous film coats (hydroxymethyl cellulose, hydroxypropyl methylcellulose),There are several typesof defects that can occur originally (orange peel effect for instance). All coatings, essentially, are such that each applied coat is not complete, so that there are overlaps, and in essence there isalways an orange peeleffect,except in a tablet this cannot beseen. It is simply assumed in this writing, that the tablets placed on stability are not defective. From a stability point of view there may be changesin appearance, mostly due to dislodging or rupture of the film. Sometimes these changes are first seen in the engraving. To properly record changes inappearance of the film, descriptive means can be used, but it is often a good idea to take a photomicrograph originally of all coated tablets (be they sugar or film coated). If defects show up in the coating as a function of time, then the question arises whether this is due to the formula (film and uncoated tablet) or to the ay in which it was made (initial defective procedure, possibly not noticeable). ost often, theseproblems result in efforts in the formulation area, and recording(visually or photographically) at many intervals (3, 6, 9, 12 months) is therefore advisable. In this manner reformulation can be carried out as soon as the problem is identified.

erty that should be monitored, both fo gloss. This is usually done subjectively. eter that assesses the gloss, but points out th re isstill a great deal tivity in the use of it. As a problem-solving tool, scanning electron microscopy is advised, because of the au mented detail it offers, a detail that often pin oints the individual some formulation setups, it is possible (e.g., with an Instron tester) to the force necessary to strip a film from a substrate. f this substrate is the tablet surface, it ispossible to evaluate films, moisture ditives, etc., to ascertain which isthe proper way in whichto r e f o ~ u l a t the e film. pearance (i.e., the color) of the film coated tablet can. bechec~ed ctance meter (or by diffuse reflectance), as described in the pressolution and disintegration are, of course, sensitive param any change in the film will be reflected in these pro~erties.

d descriptions of sugar coating procedures are beyond the scope of this in a coating pan) In brief, in sugar coated tablets there is a~plied (~sually first a barrier coat (frequently shellacor other polymer), then a subcoat (fre ), then. a dye coat consisti in^ terra alba/gelatin, with talc used as a consp usually of sucrose syrup and lake dye), then a ing coat (usually sugar syr and finally a polish coat (usually beeswax either dry or in solvent solution). carried out in a canvas coated coating pan. e typical defects onstorage are chipped tablets and ta y. The former can be tested for by using a friabilato a correlation with an actual ship test should be atte test, ta~letsare sent by various es (rail, truck, air) destinations and then back again. In so doing, it is possibl 1 stress test is comparable to the actual tablets split in the periphery it is usuallyd ot have been quite dry at the time one is due to an improperly applied barrier coat. gain, photomicrography (and in probl opy) is advocated as a reference for ch nce (i.e., the color) of the coat can be c r by diffuse reflectance), as described in the previous section. issolution and disinte~rationare, of course, sensitive parameters, because any change in the film will reflect in these pro~erties.

enteric coat is an attempt “to administer two dosesin one tablet.’, cing an acid resistant film (e.g.,a polymer containing a carboxyl g of 4 6 ) on an uncoated tablet and then sugar CO it. The first dose is contaille the core, and the seconddoseisapplied in the coat, whichshouldrelease material imme~iately.

nteric coating is a delicate operation, and often there is, in the production that in-vitro dissolution must be carried out after .eighth coat may then be applied or not depen~in outcome of the in- o test. This latter is usually the USP test that calls fo

tablets have disintegrated at the res scribe ble, but if one or two tablets fail, then it encies to polymerize.

The behavior at accelerated temperatures is not necessarily indicative of (nor extrapolable to) room-temperature characteristics.

y (198Qa) have demonstrated how, depending on the sorption pule fill, a capsule shell can lose moisture to the capsule fill conversely under opposite sorption isotherm conditions of the fill and become soft. ier, dissolution rate of the gelatin lution on storage, but gastric juice c well eliminat~such a problem. A thorough review of the p he occurrence of pellicule formation has been discus

t (1987) have studied their considerations. of gelatin at time t ,

swelling of gelatin (Fig. 35) and have grams of aqueous buffer solution is

he effect of additives can then be stud t behavior of gelatin was experienced withdifferent dr rful preformulation tool (when combined with data the drug, as de~onstratede.g. in Example 10.3). reported on the moisture isotherms of ge the powder mixture in the gelatin, it is in the previous section deali with tablets) to calculate the shift in moisture from owder ~ i x t u r e(or

f 0.0

Time, hours Swelling isotherms of gelatin.

icrocapsules may decompose as a function of time. This bas been reported by akine et al. (1987) for the case of poly(L-lactide) microcapsules. Logical means of e~timatingthe loss of intact polymer is (a) from the decrease in weight-averaged molecular weight,(b) by monitoring the loss in weight of polymer by gelpermeation chromatography, and (c) bydetermining the a t of lactic acid formed. is an example of the decrease in weight-ave olecular weight up0

uidelines advocate exposure of dosa light, and although this might be instructive, it does not represent a test that simulates conditions in actual commerce (in general).There are exceptions: certain products are liable to be kept in handbags and kept out in the open, but these are the exception. In general products are considered to be kept in controlled plant environments, in warehouses or in controlled pharmacy conditions or in (short) transit. To define a storage condition it necess is e actual conditions in (1939), Lachman and the marketplace, and this has been done by Es Cooper (1959a, 1959b), and Lachman et al. ermined the spectral c o ~ ~ o s i t i oofn light and light intensity in the typical ~ m e r i c a npharmacy, and in general it isassumed that the average foot-ca~dlesin. a pharmacyis 5-15, and 10 is used as an average. ne could now proceedby checking a product for three years under suchconditions, but rather than do that, it is desirable to accelerate the conditions so as to obtain an answer somewhat more rapidly. The guidelines’ suggestion of using more energetic (UV) light is not good for such acceleratory attempts, because the more energetic light will (or may) give rise to reactions that would never take place in the light in a pharmacy (which is much more poor in ultraviolet light). lamp 1.5”in diamLachman and Cooper determined that a #48 12 C eter and 48” long produced a good averagespectrum and produced 3250 lumens per

14 12

4 0

100

150

Time (days) eight-averaged molecular weightof poly(b1actide) microcapsules. (Figure constructed from data reported by Makine et al., 198’7.) Appearance Change of a Tablet as a Function of Time in a Lachman-Cooper Light Cabinet Storage Amber dishOpen (weeks) 0 4 8 12

bottle 0.302 0.212 0.188 0.190

0.302 0.315 0.312 0.309

60 watts. They suggested acceleratingthe test by increasing the lumen reachedby the dosage form, and they increasedthis by placing it closer to the light source. The light quanta absorbed by the dosage form are inversely proportional to distance (d)of the dosage form from the light source squared, i.e., proportional to 1ld2.They describea light cabinet, equipped with shelves closeto the described light source.,adequately ventilated so that the temperature does not rise substantially. Lightmeters on the shelves allow either movement of the shelf so that the light intensity is always the same, or simply servesas adevice to control when the bulbs should be changed. If used as described, 1 month in the light cabinet is equivalent to 24 months on A typical set of data of the discoloration of a the average pharmacist’s average shelf. dyed tablet (as gauged by optical density) is shown in Table 8. It is seen that the cabinet allows for evaluation of the means of preventing photoinduced changes or reactions to take place (useof the amber bottle). This could be accomplished byother means as well (coating of a tablet, using an opaque rather than a transparent capsule., using a capsule with a dye that screens out the part of the spectrum that causes the photoreaction). In the case of uncoated tablets, most often the dye will discoloror fade in the upper layer only. If the tablet is broken, the color is still intact on the inside. If it fades all the way through, then this is indicative of photoinduced interaction. In the case of ethinyl estradiol and an F C dye, Kaminski et al. (1979) showed

interaction, but this interaction wasfirstnoticed in photocatalyzed. lations on light testing suggest the use of xeno ight is applied ina very short period of time ( rm (and the drug and the drug product) is re of this with “real time and exposure” is unknown, and a re it can be rationally analyzed.

here is very little in literature regarding the sta apers, the reagent is adsorbed on filter by a ~angmuirisotherm; there will, h ese may bind part (often a ence there will bean initial “loss” of reagent, an for by excesses, since only the reagent that is not c reaction in the ~iagnosis. tability data are gauged on the basis of initial assay ( Usually the stability is evaluated in a semi~ua~titative i.e., an operator will carry out the ~iagnosisinitially. that monitors sugar content (e.g., in urine), the filter p sucrose solutions of various concentrations. The instr reaction is “positive” if a certain color is achieved (i.e.,the concentration of s~crose is above a level, L negative if the responselevelisbelow ther c o n ~ ~ ~ t r a t i o n , L*, where L* -cL area in betweenis then 66doubtful,”a retest. Initially t shouldevoke the correct response ( s ~ ~ a b l quality y, control released). sition, which will vary from strip to S, q, maygive an incorrect respon ise, is this p a r a ~ e t e rq, which ion date with 95% confidence that q is below 5%.

It was shown inthe chapter dealing withstatisti are mathematical means of calculating expir dat is not directly possible with physical the lty that exists in ~uantitatingthe p ‘the physicochemical c h ~ n g or other external in~uence,sh efficacy of the product.’’ This ehow to transform the experi~entaldata

,a suspension may start c en up in a reasona~lelength a criterion must be set by the investi tory group within an or~anization. everal containers of di

e, then it should be

evaluated, and the “worst acceptable” (akin to the lower acceptable quality limit in The investigator now can do a rotation test on this oore and Lemberger (1963), and state e.g. that after 20 controlled rotation§, an assay of the s~pernatantmust not be lessthan L mglmL. e of test can then be carried out at various storage times at room ure; L can be plotted versus time, and a usual statistical test performed n other words, for physical testing that has no number associate rtant to a t t e m ~ tto find such a number. n some cases this is not possible. A case in point is the particle size of an intra~enousoil emul§io~. ~oalescence and formation of free oil will result in toxic t is next to impossible to determine an acce table upper limit

Carstensen, J. T.,

ev. Ind. Pharm. 19:1811.

Carstensen, J. T., Wright, J. L., Blessel, K., Sheridan, J. (1978a). J.Pharm. Sci. 67:48. Carstensen, J. T., Wright, J. L., Blessel, K., Sheridan, J. (1978b). J. Pharm. Sci. 67:982. Carstensen, J. T., Kothari, R,, Chowhan, Z.T. (1980a). Prod. Dev. Ind. Pham. 6569. Carstensen, J. T., Kothari, R., Prasad, V. K., Sheridan, J. (1980b). J. Pharm, Sci. 69:290. Carstensen. J. T., Alcorn, G. J., Hussain, S. A., Zoglio, M. A. (1985). J. Pharm. Sci. 74:1239. Carstenen, J. T., Franchini, M., Ertel, K. (1992). J. Pharm. Sci. 81:303. Chafetz, L., Hong, W.-H., Tsilifonis, D. C, Taylor, A. K. Philip, J. (198 73: 1186. Chowhan, Z. T. (1979). Drug Dev. Ind. Pharm. 541. Couvreur, P. (1975). Thesis.Docteur en Sciences Pharmaceutiques. Louvain, Belgium: IJniv. Catholique, p. 87. Cruaud, O., Duchene, D., Puisieux, F., Carstensen, J. T. (1980). J. Pham. Sci. 69:607. Davis, S. S. (1984). In: Asche, H., Essig, D., Schmidt, P. C. eds. Technololgie von Salben, Suspensionen und Emulsionen. Stuttgart: Wissenschaftliche Verlag~gesellschaft,pp. 160-175. avis, S. S. (1987). In: G r i m , W.,ed. Stability Testing of rug Products. Stuttgart: Wissenschaftliche Verlagsgesellschaft, p. 40. avis, S. S., Khanderia, M, S,, Adams, I., Colley, I. R., Camrnack, J. Sanford, T. J. (1977) Texture Studies 8:61. .G., Smith, D,, Weierstall, Dey, M., Enever, R., Kraml, 10:1295. Digenis, C. A., Gold, T. B. Shah, V. P. (1994). J. Pharm. Sci.83:915. Dukes, G. R. (1984). Drug Dev. Lnd, Pharm. 10:1413.

Cheng, H. (1993). J. Pham. Sci. 82:l 1085. Lieberman, H. A., Mod ,S., Ninger, F. C. (1967). J. Pharm. Sci. 56:63. udraraju, V.S., Fhie, J. Chowhan, Z.T. (1993). Int. J. Pha~maceutics 97:119.

.V. (1969). Manuf. Chem. Aerosol News. 40:29. Grimm, W., Schepky,G. (1980a). Stabilitatsprufu~gin der Pharmazie. Aulendorf, Germally: 80b) Stabilitatsprufung in der Pharmazie. Aulendorf,

erm many:

Grimm, W., Schepky, G. (1980~). Stabilitatsprufung in der Pharmazie, Aulendorf, Germany: Editio Cantor, p. 230. Grimm, W., Schepky,G. (1980d). Stabilitatsprufung in der Pharmazie. Aulendorf, Germany: Editio Cantor, p. 216. Gucluyildiz, H., Banker, G. D., Peck, G. E. (1977). J. Pharm. Sci. 66:407. ahn, A. U., Mittal, L. (1979). Colloid Polym.Sci.257:959. y, S. A.(1971). Pharmazie 26:636. ammouda, Y .,Sala arris, R. P.(1968). Aerosol Age 18(1):36. Higuchi, T. (1963). J. Pharm, Sci. 52:1145. ard, B.(1983). J. Pharm. Pharmacol. 34:610. Cosmet. Ind, 107( 1):46 The Aerosol Handbook. Caldwell, NJ: Dorland. Jtargensen, K., Christensen, F. N. (1996). Int. J. Pharm. 143:223.

J0rgensen, K., Christensen, F. N. Jacobsen, L. (1997). Int. J. Pharm. 153:l. .C. (1969a). Mfg. Chem. 40(8):63. .C. (1969b). Labo-Pharma. 17(9):71. Kaminski, E. E., Cohn, R. M., McGuire,J. L., Carstensen, J. T.(1979). J. Pharm. Sci. 68:368. Katori, N., Okudaira, K., Aoyagi, N., Takeda, Y., Uchiyama, M. (1991). J. Pharmacobio-Dyn.14:567. Kokobo, T., Sugibayashi,K.,orimoto,Y.(1991). J. Control Release 17:69. Kokobo,T., Sugibayashi, K., orimoto, Y. (1994). Pharm. Res.11:104. Komerup, A., Wanscher, J. H. (1967). The Methuen Book of Colour. London: met hue^. Lachman, L., Cooper, J. (1959a). J. Am. Pharm. Assoc. Sei, Ed. 48:226. Lachman, L., Cooper, J. (1959b). J. Am. Pharm. Assoc. Sci. Ed. 48:233. Lachman, L,, Schwartz, C. J., Cooper, J. (1960). J. Am. Pharm. Assoc. Sci. Ed. 493226. Lee, S., DeKay, H. 6.,Banker, 6. S. (1965). J. Pharm. Sci. 54:1153. Lee, TI-"., Notari, R. E. (1987). Pharm. Res. 4:98. Leu, R., Leuenberger, H. (1993). Int. J. Pharm. 90:213. Leuenberger, H., Leu, R. (1992). J. Pharm. Sci. 81:976. Lik, J., Cadwell, K. D., Anderson, B. D. (1 993). Pharm. Res. 10:535. Makine, K., Ohshima, H., Kondo, T. (1987). Pharm. Research 4:62. Matthews, B, A.9 Matsumoto, S., Shibata, M. (1974/75). Drug Dev. Comm. 1:303. McGinity, J. W. (1993). Pharm. Res. 10:542. McNamara, P. J., Foster, T. S., Digenis, 6. A., Patel, R. B., Craig, W. A., Welling, P. C., Rapaka, R. S., Prasad, V, K., Shah, V. P. (1987). Pharm. Research 4:150. cVean, D., Tuerck, P., Christenson, C., Carstensen, J. T. (1972). J. Pharm. Sci. 61:1609. ehdizadeh, M,, Crant, D. J. W. (1984). J. Pharm. Sci. 73:1195. Mendenhall, D. W. (1984). Drug Dev. Ind. Pharm. 10:1297. Moore, A., Lemberger, A. P. (1963). J. Pharm. Sci. 52:223. Muller, R., Bardon, J., Arnaud, Y., Champy, J., Roland, P. (1979). Ann. Pharm. Fr. 37:301. Murthy, K. S., 6hebre Sellassie, I. (1993). J. Pharm. Sci. 82:113. Nakagawa, T,,Uno, T,(1978). J. Pharmacobio-Dyn. 6:547. Nakamura, A. Okada, R. (1976). Colloid Polym. Sci. 254:718. Naylor, L. J., Bakatselou, V. ,J. B. (1993). Pharm. Res. 10:865. Nogami, H., Nagai, T., Uchi ).Chem. Pharm. Bull. 14:152. (1987). J. Pharm. Sci. 76:715. Omelczuk, M. O., McCinity, J. W, (1993). Pharm. Research. 10:542. Ondracek, J., Boller, J., Zullinger, F. H., Niederer, R. R. (1985). Acta Pharm. Technol. 31:42. Pengilly, R. W., Keiner, J. P. (1977). J. Soc. Cosm. Chem. 28:641. Pietcsh, W. B. (1969). Eng. for~Ind.435. Pitkin, C., Carstensen, J. T. (1973). J. Pharm. Sci. 62:1215. Podzeck, F. (1993). Int. J. Pharmaceutics 97:93. Polli, 6. P. Crimm, W. M., Bacher, F. A., Yunker, M. H. (1969). J. Pharm. Sci. 58:484. Prandit, J. K., Wahi, A. K., Wahi, S. P., Mishr, B., Tripathi, M. K. (1994). Drug Dev. Ind. Pharm. 20389. Rassing, J., Atwood, D. (1983). Int. J. Pharm. 13:47. Reng, A. K. (1984). In: Asche, H., Essig, D.9Schmidt, P. C., eds. Technololgie von Salben, Suspensionen und Emulsionen. Stuttgart: Wissenschaftliche Verlagsgesellschaft,p. 203. Rhamblhau, D., Phadke, D. S., Doerle, A. K. (1977). J. Soc. Cosmet, Chem. 28: 183. Rhodes, C. T. (1979a). In: Banker, 6. S., Rhodes, C. T., eds. Modern Pharmaceutics. New York: Marcel Dekker, p. 329. Rhodes, C, T. (1979b). Drug Dev. Ind. Pharm. 5:573. Richman, M. D., Shangraw, R. F. (1966). Aerosol Age 11(5):36. Rodrigue~-Hornedo, N.,Carstensen, J. T. (1985). J. Pharm. Sci. 74:1322.

(1974). Dtsch. Apoth. Ztg. 114:1653. (1965). J.Pharm. Sci.54:260. (1983). Pharmacy International 4:225, and 173. owe, R. C. (1985). Pharmacy International 6:225. .,Halterlein, L, J., Blanchard, J. (1985). Int. J.Pharm. 26: 165. .(1974). Pharm. Ind. 36:327. W., Palermo, P. J.,Pollard-Walker, S. (1984). J.Pharm. Sci. 73:126. Sciarra, J,J.(1974). J.Pharm. Sci. 63:260. (1967). . AerosolAge12(2):65;12(3):45;12(4):65. Sciarra J. J ,Goudie, A,J., Huetteman, A. J.(1960). J.Am. Pharm. Assoc. Sci. Ed. 49:467. Lieberman, H., Chow, F.S. (1963). J.Pharm. Sci. 52:994. unzel, R.(1959). Pharm. Ind. 21:417. Shafer, E. G. E., Wollish, E. G., Engel, C. E. (1956). J.Am. P h a r ~Assoc. . Sci. Ed. 45~114. Sherman, J.(1955). Research 8:396. 4). J.Pharm. Pharmacol. 16:l b, N.(1966). J.Pharm. Pharmacol. 18:11’75. 1955). A Study of Some Emulsi~ersfor Pharmace~ticalEmulsions. PhD. thesis, Univ. of Florida. .,Bryant, S,, Welch, S., Digenis, G. A. (1984). J.Pharm. Sci. C. Caraballo, I., Millan, M., Pinero, R. T., Melgazo, L. J., (1998). Int. J.Pharm. 174:63. awara, S., Imai, T., Otagiri, M. (1994). Pharm. Res. 11:272. Suryanarayanan, R.,Mitchell, A. G. (1981). J Pharm. Pharmacol. 33:112 Suryanarayanan, R., Mitchell, A. G. (1984). J.Pharm. Sci. 73:78. J.E. (1964). J.Pharm. Sci, 53:995. (1957). Trans. Inst. Chem. Eng. 35:258. adding, H. V., Tausenfreund. R. A., ichaelis A, F.(1972). J.Pharm. Sci.61:1811. Van den Tempel,A. (1953). Stability of Oil-in-Water Emulsions. alton C. A.,Pilpel, N. (1972). J.Pharm. Pharmacol. 24:llOP. .H. (1921). Phys. Rev. 17:273.

~ ~ a r ~ a c ~ u consult t i c a l an^, Raleigh, ~ o r t Carolina h

Trigen ~aboratories, Salisb~ry, ~aryland

331 331 332 333 333 334 334 335 335 335 336 336 337 337 338 338 338

342

Oil

10.1.Keyvariables-resolution equation parameters 10.2. Isocratic or gradient mode 10.3. Role of p 10.4.Roleofsolvent type 10.5.Roleofmobilephase 10.6. Role of buffer 10.7.Roleof the ion-pair reagent 10.8.Roleof the column 10.9.Roleof temperature 10.10,Roleofflow rate

344 344 344 346 346 347 347 349 350 350

11. Optimization (Optimizing the Separation) 11.1. Peak area or peak height for quantitation l 1.2.Plackett-Burmandesign

351 351 351

12.Computer

351

Software for Method

13.Other Applications 13.1. Analytical method for cleaningassessment 13.2. Physicochemical characterization method (dissolution method) 13.3. onc chromatographic methods

352 352 352 353

latory and Compendial Basis of 353 15. Validation Protocol 16. Validation Parameters 16.1.USPGeneralChapter1225>, ethods 16.2. ICH Guidelines 16.3. FDA ReviewerGuidance

354 356 Validation of Cornpendial 356 358 359

17.Definitionof Validation Parameters 17.1. Accuracy 17.2. Precision 17.3. Specificity/ selectivity 17.4. Forced degradation 17.5. Detection limit (DL) 17.6. Quantitation limit (QL) 17.7. Linearity Range 17.8. 17.9. Robustness 17.10. Application of Plackett- urman design to ruggedness testing 17.1l . Stability of sample and standard solutions 17.12.System suitability specificationsand tests

360 361 361 362 364 364 365 366 367 368 368 369 370

18. Post Validation Issues 18.1. After the laboratory work 18.2. Revalidation 18.3.Method transfer

372 373 374 375

plication of Validation

ther Analytical Techniques

.l. Cleaning method ysicochemical characterization method (dissolution) nchromatographic methods era1 considerations

376 376 378 378 380

eferences

381

Appendices

384

According to the regulatory definition (l), a stability-indicating method is one of a number of ~uantitativeanalytical methods thatare basedon the characteristic structural, chemical, or biological properties of each active ingredient of a drug product and that will distinguish. each active ingredient from its degradation products so that the active ingredient content can be accurately measured.

Therefore a stability-indicatingmethod is an analytical procedure that is capable of discriminating between the major active (intact) pharmaceutical ingredient (API) from anydegradation (decomposition)product(s) formed under definedstorage conditions during the stability evaluation period. In addition, it must also be sufficiently sensitive to detect and quantify one or more degradation products. A corollary may be added that the analytical method must be also capable of sep any other potential interfering peals suchas an internal standard then, the discriminating “nature” of the method indicates the method to be stability-indic~tingas well as s t ~ b i l i t y - s ~ e c l ~Later c . in the discussion wewill see that other methods maybe stability-specificbut not stability-indicating.Stressed testing maybeused(1,2) to expedite e decompositionpathway@) to generate decomposition product(s) for the AFI. wever, stressed testing under forced conditions of oxidation, photolysis, hydro1 ,and varying pH values may form some decomposition products thatare unlikely to beformedunder accelerated or long-term stability storage conditions. The products generated nonetheless may beusefulindeveloping and validating a suitable stability-indicating analytical method for the analysis of the drug substance and the drug product, expediting the availability of the completed analytical method. It is paramount that thechosen analytical method usedfor stability evaluation be validated and discriminating to ensure efficacyof the subsequent stability evaluation. Confidence in the stability data is predicative on time invested up front to ensure a viable procedure as well as to conform to legal and regulatory requirements (2).

evelopment of a stability-indicatingmethod should be predicated on the method’s intended application as well as selecting a suitable technique designed to assess

19s stability requirements. Obviously the intended application of a stability-indicatingmethod is for monitoring the stability of a given drug in a finished would require assessment of the method’s stability-indicating prospecialty application of a stability-indicating method is leaning validation testing, whichwould require assessment of its stability-indicat properties, as holdingtime(of the swaps)wouldbe a cri applications such as product release, performance testing (i.e., and in-processtesting do not require this assessme e p h a r ~ a c e ~ t i chouses al still (but now commonly less due to ces and overall ind~stry UV, for p r o d ~ crelease, t practice) utilize a non-stability-indica LC method for stability te r, whenever there is a hold time issue, common in dissolution or inthe method for its stability-indicati Other chromatogra~hicsepar thin-layer chromatogra asin~ly)capillary electrop specific methods. or coupled with ion-suppress ion^ which accounts for 85Y0 or more of the general pharmaceutical applications. ~ o n c h r o m a t o g r a p ~and c spectroscopic techniques suchas titrimetry, atomic absorption9 UV spectrophotometry, and infrared spectroscopy, while precise, are not considered stability-indicating, and as such not suitable for stability assessment applications.

efore beginning with actual e~peri~entation it would be advantageous to view me tho^ development from a broader perspective. he method development process can be visualizedfrom a high-level processmap perspective better to define the general steps encountered to achieving the end product, a stability-indicating method.

pproach method development withthe intention of us' bility assessment as a final application, after the roach entails determining the discriminatin front before investing timeand money in evaluating other analytical parameters prior to assessin the stability-indicating element of the method. G is the method of choice for stability-in ' , although thin-layer chromatography , and capillary electrophoresis (CE) are als ith ionic suppressionaccount ies for small molecularwe suited for applications in r ditional applications may be in cleani er techniques such as titration and release testing, are generally considere nons ecific and thus are not considered for stability assessment. ed with findingor developing a method, one or two on the nature of the chemical entity: modification used whenthere is i n f o ~ a t i o n or a method already ase, the existing method is modifiedor tweaked to 1s may or may not be suitable; if not, de~elopment he goals of the separation should also

roperties of the APZ is invaluable to the method n on the various properties has been collected, either through a systematic program of generating the appropriate infor~ation in support of drug discovery (organic chemistry synthesis) on the one hand, or on the other, from a search of the literature or from company drug profiles, spectral libraries, or reports. Information such as dissociation constants, partition coefficients, fluorescent properties (if any), chromatographic behavior, spectrophoto~etric properties, oxidation-reduction potentials, for~ulationstability studies, and solubility studies are all very useful and can expedite the develo~ment issociation constant and partition coefficients can be used to develop efficient liquid/liquid extraction procedures, and data on fluo~escence,spectrophoto~etric, chromatographic, and oxidation-reduction properties can beused to determine the best meansof meas~ringand quantifying the analyte of interest. Stability studies are erf formed on the drug substance, in solution and mixed with phar~aceutical excipients as part of compatibility studies. Labile functional groups are identified, and the susce~tibilityof the drug to hydrolysis, oxidation, thermal degradation, etc. is determined. Gompatibility studies are performed to assess the stability of the when mixed with common excipients and lubricants as well as to determine any interaction between the drug and the (inactive) rawmaterials. Solubilities should

be determined in a number of solvents covering a range of polarities that are commonly used in method development. olubilities should be determined in aqueous and organic solvents, such as Aqueous

O~ganic

Ethanol /methanol Chloroform Cyclohexane Acetonitrile Tetrahydrofuran pectral libraries are established, and information gleanedisuseful for LC separation. On the other hand, however, selection of initial conditions for an sometimes this physicochemical information may not be known or available, so that an initial separation would have to be tried, based on prior experience, in order to determine a course of action for subsequent experi~e~tation.

Ideally, knowledge of the API’s nature relative to composition and other properties would be beneficial. For example, information about the compound’s synthetic route would shed light on any related product(s) and possible degradation product(s), as well as possible impurities; knowled~eof the compound’s chemicalstructure would reveal any possible stereoisomer which in turn would necessitate a different separation strategy, and so forth. able l shows typicalinformation that would be helpful concerningthe nature of the compound. The more infor~ationis available, the less empiricalthe approach to developing a separation method will be.

To determine the separation goals, which should be clearly defined, a number of questions should be asked to help delineate the end purpose of the separation. Typical questions may include hat is the overall purpose of the eth hod-quantitativ~, ~ualitative,or for isolation/purification of a compound (i.e., content assay, stability, impurities, cleaning assay, or for purification application)? Useful Physicochemical/Related Information Concerning the Compound Wavelength of absorption (Anlax) Identity/number of compounds present (i.e., stereoiso~ers/c~iral centers?) Chemical structure (functionality); amphoteric olecular weight pK, values of compounds Salt form of the drug Solubility of compound Purity of compound

at level of accuracy and precision would be needed? e method is designed for what type of matrix? ow many types of sample matrices are encountered? Is the method developed using certain equipment transferable to the control laboratory, which may not have the same equipment? ill the method be used for a few samples or many Sam raphic parameters are needed? olution is needed? at is a suitable/acceptable separation time? is a suitable column pressure? much sensitivity is required? S an internal standard needed? re there any detection issues? ost analytes absorb in the UV region of the oes integration use peak area or height? Is the mode isocratic or gradient?

hile there are a number of LC methods available to the development chemist, perhaps the mostcommonlyapplied method isreversed-phase.Reversed-phase and reversed-phasecoupledwith ion-pairing probably account for more than 85% of the applications for a typical pharmaceutical compound. The typical pharmaceutical compound is considered to be an API of less than 1,000 daltons, either solublein water or in an organic solvent. The water-soluble API is further differentiated as ionic or nonionicwhich can be separated by reversed-phase. Similarly, the organic soluble A I can be classedas polar and nonpolar and equally separated by reversed-phase. In some cases, the non-polar API may have to be separated using adsorption or normal phase HPLC, in which case the mobile phase would be a nonpolar orga nt. For those “special” compounds that do 000 daltons [biopharmaceuticals], isom not fall into this category enantiomers), other chromat c modesmay benecessary for separation. include ion-exchan e and chiral chromatography. In this discussion of developing a stability-indicatin LC method, onlyreversed-phasewill be discussed.

Thus,given the limited number of methods with stability-indicating properties, it is probable that the method selected wouldbe HPLC. Two veryadvantageous characteristics of HPLC, its discriminating power and its ability to operate at room temperature or at low elevated temperature, would not contribute to the degradation of the analyte. It is further assumed that the API is of low molecular weight (3 batches to describe the manufacturing variability. Shelf life specifications describe the quality at the end of the shelf life and include Tolerable changes during storage and shipment. Therefore corresponding stability data are required with organoleptical, physicochemical, chemical, and microbial tolerable changes. iti

A basic distinction is drawn between stress, accelerated, and long-term storage conditions; see Table 3. Stress, Accelerated, and Long-Term Storage Conditions Type

Condition

Stress

Temperature: 10°C higher than accelerated temperature of 40"C, e.g., 50"C, 60°C, 70°C 2 -10°C Temperature cycle 5 4 0 ° C Open storage at 25"C/60% r.h., 3OoC/7O% r.h. and 40°C/75Yo r.h. Xenon lamp 48 hours

Accelerated

4OoC/75% r.h. (3ooC/~O~0 r.h.)

Long-term

25"C/60% r.h. 3O0C/7O% r.h.

The conditions used in stress and accelerated tests are above those of the relevant clibatic zones, allowing The discriminatory power of the analytical procedure to be verified. Weaknesses of a formulation to be identified. Stability information to be generated. The last-named aspect is particularly important in designing a stability program for clinical samples. To ensure ~ontinuousdevelopment, specific stress and acceleration tests are carried out, which are then verified by long-term tests.

f stress and acceleration tests are to be successful, two aspects must be paid special attention: Clear se~arationbetween the tests for organoleptic and physicochemical stability chemical/microbial stability Use of packaging materials i m p e ~ e a b l eto water vapor for stress tests at elevated temperatures, allowing application of the laws of reaction kinetics The reaction mechanis~may change with higher temperature if the samples are dried. It is therefore necessary to use packaging materials that are impermeable to water vapor to prevent solidformulations becoming dehydrated at higher temperatures or the active ingredient concentration of liquid formulations increasing due to loss of moisture. he laws of reaction kinetics cannot be used to make stability predictions for leptic and physicochemical changes. his is the reason that solid dosage forms, for example, are stored without packaging at ~5"C/60%r.h. This induces the maximum possible changes due to absorption or loss of water. Semisolid and liquiddosage forms are stored at 2 -lO"C, semiliquid at 540°C in order to detect irreversible changes. Storage at 30°~/70%r.h. is not usually necessary because most clinical trials are performed in countries of climatic zones I and 11,

The testing frequency is established to suit the problem beingstudied. Retest periods are different for stress, acceleration, and long-term tests.

ith all the different strengths, dosage forms, and packaging materials examined pment phase, it is not possible to provide three batches for each liable information can be obtained nevertheless by applying

Storage period ( ~ o n t h s )

erated term Stress phase Clinical I

I1 I11

3 3

1.5 3.0 6 6

6 12-18 24-36

rim

scientifically basedrationalization measures. This includes the expedients mentioned in the ICH Stability Guideline: Bracketing Matrixing Both methods are based on the assumption that a reduced number of investigated samples is representative of the stability behavior of all samples. In bracketing, only “limit samples” are tested, for example: the lowest and highest dosage, the smallest and largest container. In matrixing, selectionisperformed according toa statistical procedure (random number). A rational bracketing system for all dosage forms would be as in Table 5. A rational matrixing system would be as in Table 6.

tested Samples Dosages 1-2 3-4

all lowest highest lowest middle highest

e6 Tests

Test sample Beginning, end Intermediate values At each testing point analyzed.

all 1/3 or 2/3 designa 1/3 or 2/3 of all samples are

At higher temperatures, desorption and loss of moisture also occurs at higher relative humidities. Unless packaging materials impermeable to water vapour are used for stress tests with solid dosage forms, the samples lose moisture at different rates in the temperature range 4O-6O0C, and the results are not suitable for a reaction kinetics calculation. Packaging materials permeable to water vapor can however also result in a falsification of the results for semisolid and liquid dosage forms if varying degrees of weight loss occur that lead to differences in the active ingredient concentration or ion strength. The use of inert standard packaging materials that are impermeable to water vapor. is thus an important precondition for stress tests that are to be evaluated in terms of reaction kinetics, and on the results ofwhich stability predictions are to be based.

An overview ofthe most important packaging materials for stability testing for clinical samples is given in Table 7.

A systematic approach should be adopted in the evaluation and presentation of the analytical results. Tests for significantchangeswith the aid of statistics, reaction kinetics calculations, or linear regression analysis are valuable tools. The stress and accelerated test results are evaluated for each clinical phase taking into account the specific objective of the respective storage conditions. On the one hand, there is the test for organoleptic and physicochemical stability, on the other hand for chemical or microbial stability. If, for example, discoloration, a decrease in hardness, an increase in dissolution, or phase separation recorded, but they is observedafter 3 months's storage at 70"C,these changes will be are only of limited relevance for predicting stability. If there are no significant changes in the test for organoleptic and physicochemical stability, stability can be predicted by means of reaction kinetics calculations that are based mainly on the results obtained after storage at stress temperatures. Considering these facts, alltest criteria can be included.in the stability prediction. A critical examination is conducted to determine whetherrelevant changes haveoccurred and whether the proposed minimum shelf lifetolerance limits have been reached or exceeded. Stability studies for clinical samplesare based on stress and acceleration tests with the aim of speeding up, especially, chemical decompositionby storing samples at elevated temperatures. The results are then used to calculate the stability behaviour at 25"C/60% r.h. based on the laws of reaction kinetics. The equations for a first-order reaction and the Arrhenius model are used. If decomposition levels are available for only one temperature ( 4 3 , the expression AE: 83 kJ mol" is used for the activation energy. Table 8 shows decomposition levels for 25"C/60% r.h. calculated from values obtained after storage at 40°C to 70°C. Reported in Table 6 are the decomposition determined after storage at accelerated and stress temperatures, and the decomposition for 25°C derived from these data. After evaluating the data of one batch it is assessed whether the strengths or the dosage forms exhibit differentstability behavior or whether the results of the batches can be combined to produce uniform stability information. The packaging material and possible interactions have to be included in the evaluation. The general stability information, the period of use, and if necessary storage instructions are based on Primary data The results of the stress and accelerated investigationsand later the results of the long-term testing for confirmation. Supportive data Drug substance stability profile, which also includes orientational predictions regarding the chemical stability of the drug substance in solid, semisolid, and liquid dosage forms.

c ij

a, U a,

+J +J

f/ do

Stability investigations during formulation findings for the clinical phases 1-111, whichprovidespecific information regarding the influenceof excipients and the overall formulation on organoleptic, physicochemical, chemical, and microbial stability. For phase I1 Minimum shelf life for clinical phase I: prediction and confirmation, For phase I11 Minimum shelf lifefor clinical phasesI and 11: prediction and confirmation, It is an important fact that the predicted minimum shelf lives will alwaysbe confirmed by the concurrently performed long-term tests. Reaction Kinetics Extrapolation Decomposition found/ extrapolated decomposition (YO)

Clinical phase Storage condition

I I I1

I1 I11

months 40°C: 1.5 25°C: 3 months months 40°C: 3 25°C: 6 months months 60°C: 3 25°C: 12 months 25°C: 18 months months 40°C: 6 25°C: 12 months 25°C: 18 months months 70°C: 3 25°C: 24 months 25°C: 36 months

0.10

to.10 0.10

tested All Highest Lowest Highest Middle Lowest

Examples: If ~ i ~ m shelf u m livesare required for 10,20,40,60,80,120 mg, then 10,40,120mg are tested. If 3 4 strengths have to be investigated, the two extremes, the highest and lowest, are fully tested. The most probable final strength, however, will be a middle strength. Therefore also a middle strength will be included with a reduced stability protocol, e.g., chemical stress testing only at 60°C. Packaging material: in selecting packaging material, the following has to be considered:

At higher temperatures, desorption and loss of moisture also occurs at higher relative humidities. Unless packaging materials impermeable to water vapor are used for stress tests with solid dosage forms, the samples lose moisture at different rates in the temperature range 40-60°C and the results are not suitable for a reaction kinetics calculation. Packaging materials permeable to water vapor can however also result in a falsification of the results for semisolid and liquid dosage forms if varyi of weight loss occur that lead to differences in the active ingredient CO or ion strength. The use of inert standard packaging materials that are impermeable to water vapor is thus an important precondition for stress tests that are to be evaluated in terms of reaction kinetics, and on the results on which stability predictions are to be based. t of the stress tests are carried out in standard packaging material. following standard packaging materials are used: Soliddosageforms:50-mLglass container withtwist-offclosure Polypropylene tube Semisolid dosage forms: Standard tube Small volumetric flask Aluminum tube, inert internal lacqu Liquid dosage forms:25mL volumetric flaskwith ground However, further investigations for the selection of the final packaging are necessary. Selection of packaging material for solid dosage forms. n the basis of the results of the stress tests for solid dosage forms, the ity to moisture can be determined and suitable packaging materials can be selected, As a rule, no interactions are to be expected. If the final packaging material has been selectedand samples packedin the final packaging material are available, the investigation of photostability should be perfo~med. Photostability: The samples with and without container are irradiated with a Xenon lamp (Atlas Suntest, 250W/m2) for 24 hours. Test criteria: appearance, drug substance decomposition and assay. Selection of packaging material for semisolid dosage forms. Suitable tests have to be carried out. ackaging: Aluminum tube internally lacquered, plastic tubes. roblems: Corrosion of metal tube; interaction with internal lacquering; sorption; perm~ationof water vapor, oxygen, aromas, essential oils. Tests packaging material-dosage form: To test for corrosion, the filled metal tubes are stored horizontally, upright, and inverted at 40°C for 3 months and are then investigated. To test for permeation and sorption, the filled plastic tubes are stored for 3 months at 50°C, 40°C, 3OoC/70%. When selectingthe packaging material, the climatic zonein which the product is to be introduced must also be taken into account.

ause of the problems arising with plastic tubes, aluminum tubes are If the final packaging material has been selected, the investigations on the performed. ith and without container are irradiated with a /m2) for 24 hours. est criteria: appearance, drug substance decomposition and assay. election of packaging material for liquid dosage forms. ackaging: ampoule, injection vial withrubber stopper, glass bottle or plastic ith screw closure or pilferproof closure and liner. leakage, desorption, sorption, permeation, interaction with action with liner. ests packaging material-dosage form: To test for sorption, permeation, p and leakage, the final formulation solution isfilled in the container, and for desorption placebo solution is used. The samples are stored vertically and inverted under the following conditions: 50°C, 40°C, 3OoC/70% for up to 12 weeks. Testing intervals: 0, 1, 2, 3 months. If the final packaging material has been selected the investigations on the photostability are performed. otostability: The samplesincoloressglass and the original packaging are irradiated with a Xenon lamp (Atlas Suntest, 250 W/m2) for 24 hours. Test criteria: appearance (colour of solution), clarity of solution, drug subdecomposition and assay. valuation: a systematic approach is adopted in the evaluation and presentation of the analytical results. Thereby all test criteria are included. ~nvestigationsfor organoleptic and physicochemical stability. here possiblethe data areevaluated for significant changes with the aid of statistics. Investigations for chemical and microbial stability. f decomposition and fall in assay has taken place, the equations for a first-order reaction and the Arrhenius equation are used. If decomposition levels are available for only one temperature, the activation energy AE: 83 KJ x mo1-I (8) is applied. The decomposition levels for the required storage temperatures are calculated. ~rganolepticand physicochemical changesthat have taken place at stress temperatures at 40-70°C are recorded, but they are only of limited relevance for predicting stability. Stability information: all results and the stability information derived therefrom are compiled in a stability report. The different stability reports are str~cturedand written in the same format. They contain aterial and methods esults and evaluation Conclusions

rimm

The stability information corresponds to the stage of d e ~ e l o ~ m e(Table nt 21). Correspondence of Stage of Development with Stability Information Stage of development

Stability information

General

Assessment of the formulation General prediction of shelf lives Storage instructions if required Proposal for packaging materials

Toxicological samples

Minimum shelf life (period of use) for toxicological samples in climatic zone I and I1

Clinical phase I

Minimum shelf life (period of use) for phase I in climatic zone I and I1 Storage instructions if required Proposal of suitable packaging materials Robustness of the formulation

Clinical phase IT

Minimum shelf life (period of use) for phase I1 in. climatic zone I and 11 Storage instructions if required Proposal of suitable packaging materials Robustness of the f o ~ u l a t i o n

Clinical phase I11

Minimum shelf life (period of use) for phase I11 in climatic zone I and IT Storage instructions if required Proposal of suitable packaging materials Robustness of the fornulation Selection of those decomposition products that may occur under accelerated and long-term testing of the registration batches in step 4 Elucidated structure and degradation pathway of the selected decomposition products Establishment of test criteria for accelerated and long-term testing of the registration batches in step 4 Registration release and shelf life specifications for registration batches in step 4 E s t a b l i s ~ e n tof storage conditions for accelerated and long-term testing in steps 4 and 5 Suitable packaging materials for the registration batches in step 4 Expected shelf lives for the registration batches in steps 4 and 5 and storage instructions if necessary Establishme~tof stability test protocol for registration batches

Registration batch

Comparison and evaluation of the results of the stress investigation with laboratory, clinical trial, and registration batches Robustness of the formulation

rage

General

4.3.1.

4.3.1.1.Solid Dosage Forms Stability test protocols (Tables 22, 23, 24) Organoleptic and Physicochemical Stress Testing Storage

Storage period testing frequency Test

25"C/6OYo 30"C /70% 40°C/75%

0, 2 weeks 0, 2 weeks 0, 2 weeks

Packaging material conditions Open container Open container Open container

criteria organoleptic, physicochemical organoleptic, physicochemical organoleptic, physicochemical

Photostability Stress Testing Storage Packaging material conditions Open container Xenon lamp 24, decomposition SuntestAtlas 250 W/m2

Storage period testing frequency Test

criteria

48 hours Appearance,

drug substance and assay

Chemical Stress Testing Packaging material Pretreatment conditions

testing frequency Test

50ml none glass container with twist-off closure or equivalent tight container 30°C/70%a or 25"C/60Yoa or 40"C/75Yoa ~~~

criteria

70°C 60"C 50"C 40°C

1, 2, 1, 2, 0, 1, 2, 1, 2,

3 months 3 months 3 months 3 months

chemical chemical all all

70°C 60"C 50"C 40°C

1, 2, l , 2, 0, 1, 2, 1, 2,

3 months 3 months 3 months 3 months

chemical chemical all all

Samples that have adsorbed the highest amount of water during open storage at 25"C/60%, 30"C/70%, 40°C/75% are used. B

4.3. 1.2. S~misolidDosage Forms Stability test protocols (Tables 25, 26, 27, 28) 4.3. l .3. Liquid Dosage Forms Stability test protocols (Tables 29, 30, 31, 32)

rim

Organoleptic and Physicochemical Stress Testing Packaging material

Storage conditions

Storage period frequency testing Test

Standard tube

2 -10°C

0, 4 weeks

organoleptic, physicochemica~

540°C in 24 hrs cycle

0, 2 weeks

organoleptic, physicochemical

40°C, tube stored vertically, with the neck of thepointing tube upwards

criteria

0, 3 months content

uniformity (homogeneous distribution ingredient, of active assay of material taken from the thebeginning, of middle, end

tube

~hotosta~ility Stress Testing Packaging material

Storage Storage condition

period

Xenon Open Atlas container

lamp Suntest 250 W/m2

frequency testing

Test criteria

24, 4 8 h

Appearance, substance drug decom~ositionand assay

7 Chemical andMicrobial StressTesting Packaging condition material Standard tube

Storage Storage

period frequency testing

50"C 40°C 30°C

Test criteria

1, 2, 3 months Organoleptic, 2, 0, 1, 3 months 3 months

chemical, microbial all all

Selection of Packaging Material Proposed packaging Composition material condition

Storage

period frequency Test testing

Metal tube

verum 40°C

stored ho~zontall~, appearance and upright, inverted

Plastic tube

placebo 50°C 40"C 3O0C/70% 40°C verum

criteria

0, 3 months Appearance, drug substance decomposition, of internal surface of for metal tube corrosion, interaction with internal lacquering 0, 1, 2, 3 months Loss mass (water in permeation), aromas, essential oils, desorption l, 2, 3 months essential oils, desorption 3 months essential oils, desorption 0, 3 months Organoleptic, loss in mass, drug substance decomposition and assay

Organoleptic and Physicochemical Stress Testing Packaging material

Storage conditions

25 mL ground5°C glass stoppered glass bottle or equivalent

Storage period frequency testing Test

2 "10°C

criteria

0, 4 weeks

organoleptic, physicochemical

0, 4 weeks

organoleptic, physicochemical

Photostability Stress Testing Packaging material

Storage

25 mL groundglass Atlas stoppered glass colorless bottle

Storage period frequency testing Test

Xenon lamp chemical, Suntest 250 W/m2

4824,

hrs

criteria Organoleptic, physicodrug substance decomposition and assay

Chemical Stress Testing Storage condition

Packaging material 25 mL groundglass stoppered glass bottle

70°C 60°C 50°C 40°C

Storage period testing frequency Test

criteria

2, 0, 1, 1,2, 1,2, 1,2,

all all all all

3 months 3 months 3 months 3 months

4.3.2. Toxicological Samples

The anticipated minimum shelf life of 12 weeks is derived from the data of accelerated testing and confirmed by storage at 25"C/60% r.h. (climatic zone 11). Stability test protocol (Table 33)

Accelerated and Confirmation Testing Batch Experimental batch laboratory

period Storage Packaging Storage material condition Standard

testing frequency Test 40"C/75% r,h. 25"C/60% r.h.

4,0, 2, 6 weeks all weeks 12

criteria all

le 33 Selection of Packaging Material Intended packaging material closure

StorageStorage Composition conditionTest frequency testing 50°C upright componentsinverted

Glass bottle with placebo andscrew or pilferproof with liner

verum

verum Ampoule

verum

placebo withInjection vial stopper rubber desorption verum

40°C 50"C, upright and inverted

of liner

0, 3 months

all

0, 1, 2, 3 months Appearance, loss in 1, 2, 3 months mass, desorption of 3components plastic months

50°C 40°C 50°C 40°C

0, 2, 3 months 2, 3 months 0, 3 months 3 months

50°C 40°C stored inverted

all all

0, 3 Organoleptic, months 3 months physicochemical,

of components

50°C 40°C stored inverted

4.3.3. Clinical Trial Samples, Phase 1-111

criteria

0, 3 months Desorption

50°C 40°C 3o0c/7O% upright and inverted

Plastic bottle with placebo closure

period

0, 3 months

rubber all

3 months

(5)

4.3.3.1. Solid Dosage Forms

4.3.3.1.1. PHASE I. The anticipated minimum shelf life (period of use) is 3 to 6 months. The minimum shelf life is derived from data of stress and accelerated testing after 6 weeks or 3 months, respectively, and then confirmed by the data of samples stored at 25"C/60% r.h. (climaticzone 11) up to 3 or 6 months. Stability test protocols (Tables 34, 35)

e 34 Organoleptic and PhysicochemicalStressTesting Batch Experimental open laboratory batch

Packaging condition material container

Storage Storage

period Test frequency testing

25"C/60% r.h.

0, 2 weeks

criteria organoleptic, physicochemical

le 35 ChemicalAccelerated and ConfirmationTesting Packaging Minimum Pre-

Storage Storage

life materia1 shelf Batch

Test period frequency testing

treatment condition 40" C

3 months none Experimental 50 mL glass container with laboratory batch twist-off closure 40°C 25"C/60%r.h. YY

criteria

0, 2, 4, 6 weeks

all

0, 2, 4, 6 weeks

all

3 months 3 months

all all

25"C/60%

none

50mL glass 6 months container with twist-off closure

none

40°C

0, 1, 2, 3 months

all

25"C/60%r.h.

40°C

0, 1, 2, 3 months

all

none none

25"C/60%r,h. 25"C/60%r.h.

6 months 6 months

all all

25"C/60%none PP tubes

, ?

PP tubes

If several strengths are to be investigated, bracketing is applied. 4.3.3.1.2, PHASE TI. The anticipated minimumshelflife (period of use)is 12-18 months. The minimum shelf life is derived from the data of stress testing after 3 months, firstly confirmedby 6 months data of 40°C and finally by 12 or 18 months data of samples stored at 25"C/60% r.h. for climatic zone II. Stability test protocols (Tables 36, 37) Organoleptic and Physicochemical Stress Testing Packaging condition Storage material

Batch Experimental open clinical batch

container

Storage period frequency Test testing

criteria

0, 2 weeks

25"C/60% r.h.

organoleptic, physicochemical

7 ChemicalStress andConfirmation Testing Storage Storage Prematerial treatment condition

Packaging Batch Experimental 50glass mL clinical twist-off with closure batches

container ,Y

2, 0, 1,

none 60°C

criteria

months

all

none

40" C

1, 2, 3, 6 months

all

25"C/60% 25"C/60%

60°C 40°C

0,2,1, 3 months 1, 2, 3, 6 months

all all

months 18 months

all all

18 12, 25"C/60% none or tubes PP 25"C/60% none 12, test packaging material YY

Test period frequency testing

r~mm

If several strengths are to be investigated, bracketing is applied. 4.3.3.1.3. PHASE 111. The anticipated minimumshelflife (period ofuse)is 2 4 3 6 months. The minimum shelf life is derived from the data of stress testing after 3 months, firstly confirmedby 6 months data of 40°Cand finally by 24 or 36 months data of samples stored at 25"C/60% (climatic zone11). Usually in phase I11 the final formulation is applied. Therefore the stability information for the registration batches has also to be considered. Stability test protocols (Tables 38, 39, 40) Organoleptic and Physicochemical Stress Testing Packaging Storage Storage period material condition testing frequency Test

Batch Clinical or physico-chemical pilot plant, final formulation

open container

25"C/60%

weeks 0, 2

criteria organoleptic,

0, 2 weeks 0, 2 weeks

3O0C/7O% 4OoC/70%

97 7,

Photostability Stress Testing Storage Storage Packaging Batch

condition material

Clinical or pilotproposed plant, packaging final for formulation

open container test or final material

period testing frequency Test

Xenon lamp substance Suntest drug Atlas 250 W/m2

criteria

24, 48 hrs Appearance,

decomposition assay

Chemical Stress and Confirmation Testing PrePackaging

Storage Storage

condition treatment material Batch Clinical 50mlor glass pilotcontainer plant with batch closure twist-off

50ml 25"C/60% none glass 12, container with twist-off closure tube or PP test packaging material

none

70"C 60"C 50°C 40°C

3OoC/70%"or 25"C/60% or 4O0C/75%

70°C 60°C 50°C 40" C

none 25"C/60% 24, 18, 12,

Test period frequency testing 3 3 3 0, 2,

l , 2, months 1,2, months 1, 2, months 1, 3, 6 months

1, 2, 3 1, 2, 3 0, 1, 2, 3

6 3, 2, 1, 18, 36 24,

months months months months months 36 months

criteria chemical chemical all all chemical chemical all all all all

The samples that have adsorbedthe highest amount of water during open storage at25"C/60%, 3OoC/70%, 40°C/75% are tested.

If more than one strength is includedin phase 111, full testing is performed with all strengths. On the base of these data itmay then be possibleto apply bracketing or matrixing with the registration batches in step 4. 4.3.3.2. S e ~ i s o lDosage i~ Forms

4.3.3.2.l. PHASE I. The anticipated minimum shelf life(period of use) is 3-6 months. The minimum shelf life is derivedfrom the data of stress and accelerated testing after 6 weeks or 3 months, respectively, and then confirmed by the data of samples stored at 25"C/60% (climatic zone 11) up to 3 or 6 months. Stability test protocols (Tables 41, 42) Organoleptic and Physicochemical Stress Testing Packaging Storage material condition

Batch Experimental laboratory

Standard tube

Storage period testing frequency Test 5°C

0, 4 weeks

criteria organoleptic, physicochemical

2 Chemical and Microbial Accelerated and Confirmation Testing Packaging Minimum material life shelf

Batch Experimental laboratory batch

Standard tube

3 months

Storage Storage Test period condition frequency testing 0, 2, 4, 6 weeks all

40°C 25"Cl60%

months 6

40°C

criteria

3 months

all

0, 1, 2, 3 months

all

months6

all

25"C160%

4.3.3.2.2. PHASE 11. The anticipated minimumshelflife (period ofuse)is 12-18 months. The minimum shelf life is derived from data of stress testing after 3 months, firstly confirmedby 6 months data of 40°C and finally by 12 or 18 months data of samples stored at 25"C/60% (climatic zone 11). Stability test protocols (Tables 43, 44)

3 Organolepticand PhysicochemicalStressTesting Batch Experimental clinical batch

Packaging material condition Standard 5°Ctube

Storage Storage

period testing frequency Test 0, 4 weeks

criteria organoleptic, physicochemical

Chemical and Microbial Stress and Confirmation Testing Packaging condition material

Batch Experimental clinical Standard batch tube

Storage Storage Test frequency testing

Standard tube Standard tube Proposed packaging material

2, 1, 50°C 40°C 25"C/60% 25"C160%

period criteria

3 months Chemical and microbial 0, l, 2, 3, 6 months all 18 12, months all 12, 18 months all

If severalstrengths are to be investigated inphase I or I ,bracketing is applied, 4.3.3.2.3. PHASE 1111. The antici~atedminimum shelf life (period ofuse)is 24-36 months. The mini mu^ shelf life is derived from data of stress testing after 3 months, firstly confirmedby 6 months data of 40°C and finally by 24 or 36 months data of samples stored at 25"C/60% (climatic zone11). ~ s u a l l yin phase IIll the final formulation is applied. Therefore the stability information for the registration batches has to be considered. tability test protocols (Tables 45, 46, 47) Organoleptic and Physicochemical Stress Testing Packaging material condition

Batch batch,

Standard Clinical 5°C pilot or plant final formulation

Storage Storage

period frequency Test testing

criteria

0, 4 weeks organoleptic, physicoche~cal

tube ?"10°C

0, 4 weeks organoleptic, physicoche~cal

540°C in 24 h cycle

0, 2 weeks organoleptic, physicochemical

40"C, tubestored vertically, withthe neck of the tube pointing upwards

0, l, 3 monthsContent unifo~ty within the tube. Assay of material taken from the beginning, middle, end of the tube

~~otostability Stress Testing Batch

Packaging material

Clinical pilot or Open container batch, plant final Proposed Suntest substance test Atlas drug or formulation packaging final for material marketing

Storage Storage period condition

testing frequency Test

Xenon lamp

24, 48h

250 W/m2 preservative

criteria Appearance decomposition and assay, assay

7' Chemical and Microbial Stress and Confirmation Testing Packaging condition material

Batch

Storage Storage Test frequency testing

Clinical Standard pilot tubeor plant final formulation

50°C

Standard tube Proposed test or final packaging material for marketing

40°C 30°C 25"C/60% 25"C/60%

period criteria

months 1,Appearance, 2, 3 microbial all all

0, l , 2, months 3, 6

months 12, 18, 24, 36 24, 18, 12,

36 months months

all all

4.3.3.3. Liquid Dosage Forms

4.3.3.3.1. PHASE I. The anticipated minimum shelf life(period of use) is 3-6 months. The minimum shelf life is derivedfrom the data of stress and accelerated testing after 6 weeks or 3 months, respectively, and then confirmed by the data of samples stored at 25"C/60% (climatic zone 11) up to 3 or 6 months. Stability test protocols (Tables 48, 49) ~~sicochemical Stress Testing Batch batch

Experimental laboratory

Packaging material 25 mL groundglass stoppered glass bottle

Storage Storage condition

period testing frequency

5°C

0, 4 weeks

Test criteria organoleptic, physicochemical

Chemical, Microbial Accelerated and Confirmation Testing Packaging Batch

material

Minimum lifeshelf

Storage Storage Test period condition frequency testing

Experimental 25 mL ground-glass months 3 40°C laboratory stoppered glass bottle batch mL 25 ground-glass months 3 25"c/60% stoppered glass bottle Proposed test months 3 25"c/60% packaging material mL 25 ground-glass months 6 stoppered glass bottle 25 mL ground-glass stoppered glass bottle Proposed test packaging material

40°C

0, 4, 2,

6 weeks

criteria all all all

0, l , 2, months 3

all

25"C/60%months 6

all

25"C/60% months

all

rimm

4.3.3.2. Phase 11

The anticipated minimum shelf life(period of use) is 12-18 months. The minimum shelf life is derived from data of stress testing after 3 months, firstly confirmed by 6 months data of 40°C and finally by 12 or 18 months data of samples stored at 25"C/60% (climatic zone 11). Stability test protocols (Tables 50, 51) Organoleptic and Physicochemical Stress Testing Storage Storage period material condition testing frequency Test

Batch Packaging

5°C

25 mL ground-glass Experimental stoppered clinical batchphysicochemicalglass bottle

criteria

0, 4 weeks

organoleptic,

Chemical, ~ i c r o b i a Stress l and C o n ~ m a t i o nTesting Storage Storage Batch

period testing frequency Test

Packaging material condition

Experimental clinical batch

25mL ground-glass stoppered glass bottle

60°C 40"C

25 mL ground-glass stoppered glass bottle Proposed test packaging material

25"C/60%

criteria

0, 1, 2, 3 months l, 2, 3, 6 months

all all

12, 18 months

all

1812,

25"c/60%

months

all

If severalstrengths are to be investigated in phase I or 11bracketing is applied. 4.3.3.3. Phase III

The anticipated minimum shelf life(period of use) is 24-36 months. The minimum shelf life is derived from data of stress testing after 3 months, firstly confirmed by 6 months data of 40°C and finally by 24 or 36 months data of samples stored at 25"C/60% (climatic zone11). Usually in phase111the final formulation is applied. Therefore the stability information for the registration batches has to be considered. Stability test protocols (Tables 52, 53, 54) 2 Organolepticand PhysicochemicalStressTesting Batch

Storage Storage Packaging material condition testing

Clinicalpilot or 25 mL ground-glass plant batches, stoppered glass bottle final formulation

2 -10°C

period frequency 0, 4 weeks

Test criteria organoleptic, physico-chemical

able 53 Photostability StressTesting Storage Storage condition

Packaging material

Batch Clinical or pilot plant batch, final formulation

25 mL groundglass stoppered colorless glass bottle Proposed test or final packaging material for marketing

period testing frequency Test

criteria

24, 48Organoleptic, h physicochemical, drug substance decomposition and assay

Xenon lamp Atlas Suntest 250 W/m2

Chemical, Microbial Stress and Confirmation Testing Storage Storage Batch

Packaging condition material

Clinical or mL,25 ground-glass stoppered pilot plant glass bottle 50°C batch, final formulation

Testperiod frequency testing

70°C 60°C 40"C

25mL ground-glass stoppered 25"C/60% glass bottle Proposed test or finalpackaging25"C/60% material for marketing

2, 0, 1, months 3 months l, 2, 3 months 1,2, 3 1, 2, 3, 6 months

criteria all all all all

12, 18, 24, 36

months

all

12, 18, 24,

36 months

all

4.3.4. Comparator or Reference Drug Products

According to the newest EU GMP Guideline (9) stability data arealso necessary for comparator or reference drug products. The following cases are differentiated: When the samples are repacked into packaging material that is as tight or tighter concerning moisture and light than the original packaging material, the original shelf life is used. The samples are repacked into packaging material that is less tight than the original packaging material. Then the samples are tested for moisture sensitivity in the open at 25"C/60% and for photostability for 24 hours with the Xenon lamp (Suntest). Test criteria: average mass and appearance. If no changes take place the original shelf life is valid. If changestake place, tighter or more protecting packaging material must be selected. Then the original shelf life is used. If samples are reworked (tablets are ground and filled into capsules), stability protocol for phase I is appliedwith the difference that the samples are stored at 25"C/60% in the intended packaging material up to 18 months for phase 11 and 36 months for phase 111; Table 55. Testing specifications:Testing specification for release and stability testing of clinical samples.

Stability Protocol for Reworked Comparator Drug Products Clinical Minimum shelf life phase ~-

12-18 months

I11

24-36 months

Storage conditions Temp. rel. hum. ("C) (%)

Packaging material

Pretreatment

Twist-off Twist-off Twist-off Proposed packaging material

25"C/60% none none

40 40 25 25

none 25"C/60% none none

40 40 25 25

Twist-off Twist-off Twist-off Proposed packaging material

none

Storage frequency Storage period

- 0 1 2 3 6 - 0 1 2 3 6 60 60

12 18 12 18

months months months months

12 12

months months months months

- 0 1 2 3 6 - 0 1 2 3 6 60 36 60 36

24 24

18 18

If stability data are notavailable and it is not intended to consider or perform further stability investigations during the clinical trial, then the minimum shelf life of the comparatoror cannot be longer than 25% of the remainingshelflife maximally 6 months whichever is shorter. 4.3.5. ~ e g i s t r a t i o nBatches

The stability information derived for a registration application should be based on the results of preclinicaldevelopment, the clinical batches and the registration batches. Thereby it can be assured that the patient after marketing authorization gets the same quality as the patient during the clinical investigation. Therefore it is important notonly to confirm the predicted quality by the data of the accelerated and long-term testing with the registration batches but to compare stress data directly. Consequently stress tests are performed also with registration batches. Usually the first registration batch is tested. The extent depends on the available stress data, but mostly a reduced stability protocol is sufficient. 4.3.5.1. Solid Dosage Forms

~tabilitytest protocols (Tables 56, 57) Organoleptic and Physicochemical Stress Testing Batch data

Packaging material

Storage Storage condition

~egistration open container fixed from available batchphysicochemicalhighest where amount was adsorbed

period testing frequency Test 0, 2 weeks

criteria organoleptic,

cal

ility

7 Chemical Stress Testing Storage condition

Storage period testing frequency

Test criteria

none

70°C 50°C

1, 2, 3 months 0, l , 2, 3 months

chemical all

30"C/70% or 25"C/60% or 4OoC/75%

70°C 50°C

1, 2, 3 months 0, 1, 2, 3 months

chemical all

Batch

Packaging material

Pretreatment

Registration batch

50 mL glass container with twist-off closure

4.3.5.2. S~misolidDosage Forms Stability test protocols (Tables 58, 59) Organoleptic and Physicochemical Stress Testing Storage Storage period condition testing frequency Test

Packaging material Batch RegistrationProposedfinal material batch packaging for marketing

hemical crobial

physicochemical cycle

2 -10°C 0, 5 4 0 ° C in 24h

criteria

4 weeks organoleptic, 0, 2 weeks organoleptic,

Chemical and Microbial Stress Testing Batch Packaging material condition

Storage Storage period testing frequency Test

criteria

2, 3 months appearance,

Registration 1,Proposed 0,50°C batch material packaging for marketing

4.3.5.3. ~ i q u i dDosage Forms Stability test protocols (Tables 60, 61) Organoleptic and Physicochemical Stress Testing Batch Packaging material condition Registration Proposed packaging batch material for marketing

Storage Storage period testing frequency Test

2 -10°C 0,

criteria

4 weeks organoleptic, physicochemical

Chemical Stress Testing Packaging material

Storage condition

Storage period testing frequency Test

criteria

Proposed packaging material for marketinga

70°C 50°C

0, 1, 2, 3 months 1, 2, 3 months

all all

atch Registration batch a If

plastic bottle, a 25 mL ground-glass stoppered glass bottle is used for chemical stress testing.

Evaluation Theresults of the stress tests are evaluated carefully.Wherescientifically justified, statistics, reaction kinetics, and linear regression analysis are applied. The analytical procedures are assessed, and then the data are evaluated as follows: rganoleptic properties Physicochemical properties Chemical (and microbial) properties ~ackagingmaterial properties obustness of the formulation The results of the stress investigations are also assessed relating to the robustness of the formulation. uring the development the formulation was challenged by the following factors: Different batches of drug substance and excipients ifferent compositions ifferent manufacturing processes with different types and sizes of equipment ifferent sites of manufacture Scale of manufacture, laboratory, pilot plant The resulting batches manufactured with all these different influencing factors have been investigated in stress and confirmation testing. Therefore considerable informatio~is available on the possibleinfluence of these factors on the stability of the drug product, and the robustness of the formulation can be evaluated. These data are an important base to assess later during running pro influence of variations and changes on the stability. Finally it is decided whether it is necessaryto ensure compliance with the minimum shelf life that has been established by providing storage instructions to be placed on the packaging material. All the results, the information on the tested batches, the applied analytical procedures, and the derived stability information, are compiled in stability reports.

All the stability reports are structured in the same format. The following stability reports are written: and Long-Term Testing with the drug product phase and Long-Term T ting with the drug product phase and Long-Termestingwith the drug product p including re~istrationbatches. The stability information of these three reports is then summarized in the tability profile of the drug product. tability information The following information is available: Assessmentof the analytical proceduresoptimized and validated during development of the drug product Establishment of the degradation pathway, under the influence of the excipients, of degradation products that may be formed under accelerated and long-term testing stablishment of shelf lives.~ i n i m u mshelf livesthat have been derivedfrom the results of stress investigations and confirmed by the data of long-term clinical phaseI: 3-6 months; clinical phase11: 12-1 8 months; clinical 11: 24-36 months. Expected preliminary shelf life for the registration batches in step 4: e.g., 24 months. Anticipated shelf lifefor the drug p in steps 4 and 5: e.g., 5 years Storage instructions: clinical phase 1-111; registration batches in step Identification of problems that could arise during storage and especially during transport. election of suitable packaging materials for clinical trial batches; for registration batches for all relevant climatic zones in step 4 Selection of test criteria for the accelerated and long-term testing with the re tration batches in step 4 Establishment of specifications for tolerable changes during storage, registration shelf life specifications Establishment of stability test protocol for the registration batches obustness of the formulation Required Capacity

Dosage form Solid §emisolid Liquid Clinical trial for solid dosage forms

equired capacity for a single analysis and docu~entation

Required capacity for an analysis in stress tests (-15% serial factor) and documentation

17 h G 0.45 weeks 12 h 20.32 weeks 8 h e 0.21 weeks 20 h 0.53 weeks

15 h 0.40 weeks 10 h G 0.27 weeks 7 h G 0.19 weeks

Time to Availability of Prediction of ~ i n i Shelf m ~ Life (Weeks) Stage of development Accelerated Stressa

I I1 1

Toxicological samples Phase I Phase 2 Phase Phase 24 I11 Registration 24 batches

Analysis Storage period Stability

12 12

Preliminary"

Final "

- 1 1 -

Total period and evaluation report

12 24

1 l

27 2

15 15

8 9 15

27 27

%Inthe clinical phases I1 and 111a preliminary prediction canbe made based onthe dataof samples stored at stress conditions.

is the central part of the stability testing. The samples are usually stored under conditions equivalent to the climatic conditions of the corresponding climatic zone, and thus binding stability statements can be derived from the results. Theseform an important part of licensing or registration documen~s.

onf firm at ion of the results of stress and accelerated testing. erivation of re-test-periods for the drug substance. etermination of the influence of batch size on the stability. erivation of holding time for the bulk drug product or intermediate stages. erivation of use life, overage, if necessary. Complete validation of the analytical procedures. ting specification for release and stability testing. ivation of the shelf lives for the final f o ~ u l a t i o nof the drug product. 5.1.1.1. ~pplicationof the Basic Principles Selection of batches and samples: representative batch for re~istrationapplication (registration batch). The batches of a minimum of pilot plant scale should beby the same synthetic route and use a method of manufacture and procedure that simulates the final process to be used on a manufacturing scale. The overall quality of the batches of drug substance laced on stability should be representative of both the quality of th material used in pre-clinical and clinical studies and the quality of material to be made on a manufacturing scale. ost important are the impurity profile, the particle size and particle size distribution, possibly surface area.

est criteria: appearance, physical criteria, decomposition and assay Analytical procedures:specific for stability testing, completely validated: Specificity, linearity, quantitation limit, accuracy, range, intermediate precision, robustness. Specifications: release specifications Storage conditions: according to the four climatic zones, the following storage conditions were established for these zones: Climatic zone

Storage conditions

1 and 11 111 and IV

25"C/6Q% r.h. 3ooC/7Q~0 r.h.

hich of these are used dependson where and how the drug substance will be stored, shipped and used. 3Q°C/70% storage only if the drug substance is applied in climatic zone 111 or IV. Testing frequency: first year every 3 months, second year every6 months, then annually. Number of batches: 3. If the information provided in the Registration Application does not contain data of production batches, information should be providedat a later stage on three early production batches manufactured for marketing. The packagingmaterial should be the same as or simulate the actual material used for storage and distribution. Small fiber drum lined with polyethylene foil for drug substance not sensitive to moisture Glass bottle lined with polyethylene foil with screwcap or twist-off for drug substance sensitive to moisture to simulate tight containers. Stability test protocols: see Tables 64, 65. Climatic Zone I1 G ICH

Batch

Packaging material

Storage condition

Storage period, up to registration

3 registration Simulating

batches

25"C/60% r.h. 0,3,6,9,12, (18) proposed 40°C/75% r.h. 1,3,6 bulk storage container

Testing frequency (months) specifications on-going (18), 24,36,48,60

Testing No:

Climatic Zone I11 and IV

Batch

Packaging material

3 registration Simulating

batches

proposed bulk storage container

Storage condition

Storage period, up to registration

3O0C170% r.h. 0,3,6,9,12,(18)

Testing frequency (months) on-going

Testing specifications

(18), 24,36,48,60

No:

valuation: the data are evaluated carefully. reaction kinetics, and linear regression a If howeverthe data show so little degradation and so little variability it is normally unnecessary to go through a formal statistical analysis but merely provide a full justification for the omission. the results obtained during the course of development are then comwith the primary accelerated and long-term stability test results. ey confirm fully the predicted data from the stress investi general stability information can be based on the following primary and ortive data and the derived stability information. rimary data and derived stability information The results of the three re~istrationbatches derived: the confirmed preliminary retest period upportive data and derived stability information The results of stress testing, the stability profile derived: retest period The results of c o ~ r ~ a t i investigation on derived: reliminary confirmation of preliminary retest period The primary and supportive data are also assessed relating to the robustness of the manufacturing process. ring the development the manufactu~ingprocess was likely challen following factors: ynthetic route that changed during develo method of manufactu~e an nt types a ~ of sizes d equipment site of manufacture, laboratory~p lant, chemic~lproduction with different types and sizes of equ scale of ~ a n u f a c t ~ and r e batch size The resulting batches ma~ufacture different factors have been investigated in stress, confirma ,and lon~-termtestin nsive information is available on the possible influence of n the stability of the drug substanc~. mally it is decided whetherit is necessaryto ensure compli~ncewith the retest period which has been established by providin be displaced on the label. All the results, the information on the tested bat l procedure, and the derived stabilit inform ability report of the investigated ccording to the two different stabilit the climatic zones I11 and TV, two separate stabili required. All the Stability reports are str~cturedand written in the same format. tability i n f o ~ a t i o nin : the dossier for re~istrationa

lication are included:

rimary and supportive stability reports with the mation

rived stability infor-

Testing specifications for stability testing ~alidationreport for the analytical procedures equired Capacity Required capacity for a time point accelerated and long-term testing and on-going equired capacity for a single analysis

Osage form

Drug substance

3 weeks

Worldwide

10h c 0.27 weeks

0.21 weeks

Analysis Testing (weeks) speci~cations including full No. of validation batches Climatic zone

Climatic zone

of test samples (weeks)

15h

1 3

0.40 weeks

Total required capacity II

I1 Worldwide Climatic zone

2 6

4 6

Worldwide 5

Stages or Bulk

eriod for I ~ t e r ~ e d i aStage te he e ~ ~ i r a t i operiod n of a production batch should be calculate^ from the date of release of the batch. The date of such e should, under normal circumsta~ces~ not excee days from the date of tion of that batch. If batches are re1 ceeding 30 days from the production date, the date of pro~uction,as defined below, should be taken as the start of the shelf life. The date of production of a batch is defined as the date that the first step is performed involving combining the active ingredient with other ingredients.

mediates are ma~ufactured

ordingly they may

be stored over a longer period of time and can be these circumstances, testing specifications

are necessary. holding time has to be determin ese data may not be necessary for the tion application; it is mainly a measure. 5.1.2.1.1. APPLICATIO~OF THE BASIC PRI~CIPLE§ election of samples: the samples should be part of representative batches (pilot plant batches). Test criteria: the criteria are investigated that are ~otentiallysusceptible to during the course of storage.

"7 BasicPrinciples andConditions frequency, Testing Production Storage in zoneclimatic

storage period Packaging/container condition (months)

I-IV

O"C/75%

intended for storage

1-11

25"C/6O% intended 25"C/60%

open for storage

0.5

30°C/70%

open intended for storage

0.5

111-IV

3O0C/7O%

Analytical procedures: the analytical procedures for the dosage forms are applied or adapted accordingly. Specifications: release specifications. For storage conditions, testing fre~uency,and storage period, see Table 67. Number of batches: 2. ackaging: the packaging should simulate the actual packaging intended for storage. The steel container may be simulated by a tight glass container, lined with the applied polyethylene foil. Evaluation: results should be within release specifications. Stability information: holding time according to storage period at 2 5 * ~ / 6 0 % or 30°~/70%. The shelf life is then calculated from the date of release of the related batch of finished product, if this release date does not exceed 30 days from the date that the intermediate is introduced into the manufacture of the finished drug the shelflifeis calculated from the product. If this 30 daylimitisexceeded, date that the intermediate is introduced into the manufacture of the finished drug product. One batch of the drug product should be produced from intermediate stored for the full holding period, and this batch should be monitored during follow-up stability testing under long-term testing conditions. S. 1.2.2. ~ o l d i n gTime for 3ulk Drug ~ r o d u ~ t

Usually the bulk drug product is not packed immediately after manufacturing or only partly or it is shipped for packaging to another site. Therefore corresponding stability investigations are necessary. These data may not be necessary for the registration application it is mainlya 5.1.2.2.1. APPLICATION OF THE BASIC

PRINCIPLES

Selection of samples: the samples should be part of representative batches (registration batches) Test criteria: the criteria are investigated that are potentially susceptible to change during the course of storage. Analytical procedures: the analytical procedures for the dosage forms are applied or adapted accordingly. Specifications: shelf life specifications.

Storage Conditions, Testing Frequency, and Storage Period

Production in period storage frequency Testing climatic Storagezone Packaging condition (months) shipment container or ~~

I-IV

40°C/75% intended

open 1-11

for storage or shipment

25"C/60% 25"C/60% intended shipment or

111-IV open

30°C/70% 3O0C/7O% shipment

for storage 3

0.5 intended for storage or

For storage conditions, testing frequency, and storage period (according to requirement), see Table 68. Number of batches 2. ackaging: the packaging should simulate the actual packaging intended for storage or shipment. The steel container may be simulated by a tight glass container, lined with the applied polyethylene foil. Evaluation: if results are out of release specificationsthe data have to be compared with those of the registration batches to investigate whetherthe stabilityis at least the same.Only under this precondition the sameshelf life can be applied. Stability information: shelflife according to registration batches.The expiration period is calculated from the date of release, and the storage period of the bulk has to be subtracted accordingly. DrugProduct

5.1.3. 5.l .3.l .

In- Use Stability The derivation of in-use stability is necessary for

Drug products that arereconstituted into ausable form before administration Drug products whose stability is jeopardized once the container is opened The in-use stability of reconstituted drug products is usually limited by the chemical stability, whereas the corresponding drug products in multiple-use containers may be limitedby the microbial stability. For both cases preliminary experiments are performed during the development stage. 5.1.3.1.1. APPLICATION OF THEBASICPRINCIPLES Selection of batches and samples: representative pilot-plant batch, usually the first one that is put on stability for registration application. Test criteria: usually the test criteria are covered by the stress, the accelerated and long-term testing and only those should be followedthat arepotentially susceptible to change due to reconstitution or opening of the container.

econstituted dry product: appearance, clarity, p , drug substance decomposition and assay, for suspension also dispe bility and particle size distribution. ultipleuse container: appearance, drug substance decom~ositionand assay, microbial preservative challenge test. nalytical procedure: the testing speci~cationfor the final formula~ionof the dosage form. ~pecifications:shelf life specifications. or storage conditions, testing frequency, and storage period, see Tables 69-72. Reconstituted Dry Product Storage condition

Testing fr~quency,storage period (weeks)

~econstitutedPowder for Injection with ~ntimicrobialPreservation Storage period

Storage condition

5°C

(days)

efore storage, treatment is simulated for a duration of up to about whereas l dosage is withdrawn daily. ~ u l t i p l eUse Container

Climatic zone 1-11 111-IV

Storage condition 25"C/60% 30°C/70%

1 month represents the period where analyses after this period.

a This

Testing frequency, storage period (months) 1" 3-5 la +5 1 dosage is withdrawn daily. N o

Injection with ~ntimicrobialPreservation zone Climatic 1-11 1~1-1~

condition Storage 25"C/60% 30"C /70%

Storage period (days) 28 28

umber of batches: 2. ackaging material: commercial packaging material.

Evaluation: the results may be part of the stability report for the registration f not, such results should be available in case the regulatory stitution store in the refrigerator 2-8'6 maximal 28 days at least 6 months 5.1.3.2.

~hQtQstability ations are undertaken as confirmatory investigations if data in the re not yet available. .l. S C ~ E R U L EOF TESTS Selection of batches and samples: representative registration batch in commercial packaging. Preferably after storage for 6 months at 40°c/75%. Test samples: oduct outside immediate packin colorless glasscontainer, tablets, capsules spread in single layer oduct in dark container as control sample. drug product in blister pack (if intended for m a r ~ e t i n ~ ) drug product in immediate packaging

enon test or corres~ondin~ vera11 illumination: 1.2millionlux integrated new ultraviolet energy not test 3 0 0 ~ 0 0 n m(

hours

22 h

5.1.3.3. A c c e l ~ r a t and e ~ ~ Q n g - ~~e s ~t i nrg ~ 5.1.3.3.1, AP~LICATIONOF THE BASIC PRINCIPLES Selection of batches and samples: representative batches for registration a (registration batch). The batches should be put on stability within *

ess: shouldmeanin~fullystimulate that whic d to large scale batchesfor marketing. The process should provide ct of the same quality intended for marketing, and meeting the same quality specifications as to be applied for release of material. atch size: two of the three at least pilot scale: one tenth that of duction or for solid dosage forms 100,000 tablets or capsules the three may be smaller, e.g.,25-50000 tablets or capsules for S ssible different batches should be applied for

of the manufacture and packaging. d be produced by a validated manufacturing process. meet the specifications required for the release of material.

Must be homogeneous and consist of random samples. The container put into storage must be representative of the batch. The samples investigated from a container must be representative of all the samples in the container at the time of analysis. esides the batches for registration application it may be necessaryto select representative batches from laboratory or scaling up. Test criteria: in stability testing, the criteria of a drug product are investigated that are potentially susceptible to change during the course of storage that are especially important for quality, safety, efficacy. Analytical procedures: the analytical procedures must be stability specific and fully validated, e.g., the following validation characteristics must be taken into account: specificity linearity for drug substance and decomposition product accuracy for drug substance and decomposition product range for drug substance and decomposition product quantitation limit 0.1% according to reporting limit intermediate precision for drug substance and decomposition product robustness Specifications:it has to be differentiated between release specifications(quality after manufacture) and shelf life specifications (quality up tothe expiration date). To fix the shelf life specifications may be not easy. possible the specifications are derived from the results of steps 1-3. t it may happen that the shelf life specificationscannot be fixed before 12or 18 months' data are available, especially for decomposition products. Storage conditions: Climatic zones I and 11: 25"C/60%, 40"C/ 75%. Ifsignificantchanges take place at 40°C/75%, storage at 3O0C/7O% is necessary. 4o0C/75%. Climatic zones I11 and IV: 30"C/70~0, It is differentiated between the stability protocol for the (climatic zone 11) and the extension to countries of the climatic zones I11 and IV with 3O0C/7O% r.h. Testing frequency: up to registration application: 0, 3, 6, 9, 12, (18) months. Storage period: the long-term testing should cover at least 12 months duration at time of submission. It is continued up to the intended shelf lifeafter registration application (ongoing stability testing). Stability test protocols: see Tables 73, 74. Climatic Zone I1 G ICH

Batch

Packaging material

3 registration Proposed

batches

Storage condition

25"C/60% r.h. packaging 4OoC/75% r.h. material for marketing

Storage period, up to registration 0, 3, 6, 9, 18, 12 24, l , 3, 6

Testing frequency (months) specifications on-going 36, 48, 60

Testing

No:

GtS

Batch 3

Packaging material

Storage condition

registration Proposed 3OoC/70% r.h. batches packaging material for marketing

Storage period, up to registration 0, 3, 6, 9, 12, 1, 3, 6

Testing frequency (months) specifications on-going

Testing

18, 24, 36, 48, 60

No:

Usually it can be predicted from the data of stress investigations whether a significantchangewill take place at the accelerated storage at 40°C/75% r.h. and correspondingly storage at 3O0C/7O% is required. If however there may be a risk samples are stored in parallel at 3O0C/7O%but only analyzed if significant change has taken place after 3 or 6 months. Number of batches: 3. If morethan one dosage isintended to be put on the market, the number of batches can be reduced by applying scientifically based rationalization measures. This includes the expedients mentioned in the ICH Stability Guideline, bracketing and matrixing. A stability protocol is elaborated by applying bracketing or matrixing to reduce the number of batches. This stability protocol may be discussed with the authorities before starting step 4. If the information provided in th tration Application does not contain data of production batches, ation should be provided at a later stage on three early production batches manufactured for marketing. Packaging material: the testing for registration application should be carried out in the final packaging proposed for marketing, or when justified in packaging that simulates the final one. A drug product is frequently marketed in several packaging materials, which may differ in both type and size. However they do not all have to be included in a stability program. This applies particularly for solid dosage forms. If information about the sorption behavior is available from the stress tests, conclusionscan be drawn through inference and by analogy. In thesecases the stability program can bereducedby applying bracketing or matrixing, as mentioned in the ICH Tripartite Stability Guideline (the study protocol maybediscussed in advancewith authorities). aging materials are used: lypropylene, composite film Plastic tubes: polypropylene, polyethylene Glass bottles Aluminum foil, aluminum blister Which of these are used depends on the stability of the drug product, the climatic zone for marketing, and the market strategy.

If the drug products are known not to be extremely sensitiveto ~ o i s t u r on e the basis of the results of stress tests, the packaging materials can be used as follows in the individual climatic zones, Table 75. Use of Packaging Materials in the Climate Zones Climatic zones Packaging material

Blisters lastic tubes Glass bottles Aluminum foil, aluminum blister

I11

ing materials: uminum tube, internally lacquered Plastic containers General Suitability of Frequently Used Packaging Materials in the Climatic Climatic zones Packaging material

Plastic tubes Plastic containers Aluminum tubes, internally lacquered

111

X X

For the packaging materials for semisolid dosageforms it may be necessary to include different sizes in the stability programme. Frequently used packaging materials: Class bottles with closure Ampoules Ampoules, glass bottles with rubber stopper Plastic containers

le 77 General Suitability of Fre~uentlyUsed Packaging Materials in the Climatic zones Climatic zones

I11

Packaging material

Plastic containers Glass bottles Ampoules

~ifferentsizes are of no importance for ampoules or glass bottles. Evaluation: a systematic approach should be adopted in the presentation and evaluation of the stability information. The long-term tdsting of the representative batches for the registration application should cover at least !:l months duration at the time of submission. xtrapolation of the shelf life up to at least 2 years may be acceptable, where supported by stress, accelerated, and real-time data. his may be especially necessary as long as the expiration date has not yet been covered by real time data for the full storage period. An acceptable approach for ~uantitativecharacteristics that are expected to decrease with time to is determine the time at which the 95% one-sided confidence limitfor the mean degradation curve intersects the acceptable lower specification limit. Statistics and reaction kinetics are also valuable approaches depending on the results. The assay, levels of degradation products, and other appropriate attributes must be included in the evaluation. Where the data show so little degradation and so little variability that is apparent from looking at the data thatthe requested shelf life will be granted, it is normally unnecessaryto go through the formal statistical analysis but only to provide a justification for the omission. All the results obtained during the course of development are then compared with the primary accelerated and long-term stability test results. If they confirmfully the predicted data from the stress investigations during development, the general stability information for the drug product will be based on the following primary and supportive data andthe derived stability information: rimary data and derived stability information: The results of the three registration batches derived: the confirmed preliminary shelf lives. Supportive data and derived stability information: The data of the drug substance stability profile derived: preliminary retest period The data of the three drug substance registration batches derived: the confirmed preliminary retest period The data of the stress and confirmation investigation during the clinical development derived: minimum shelf life (period of use) predicted and confirmed for the clinical trial phases 1-111 The data of the stress investigations with the final formulation including one registration batch derived: preliminary shelf lives for registration batches The primary data are also assessed relating to the robustness of the formulation.

It is challenged by different batches of drug substance and excipients.

Finally it is decided whether it is necessary to ensure compliance with the expiration date established by providing storage instructions to be displayed on the pack. torage instructions to be displayed on packs must reflect a genuine necessity, must takeinto account the prevailing temperatures of the respective climatic zone, and must be derived directly from tests. A logical relationship to the storage conditions thereby results, For Japan, drug products shown to be stable up to 3 years need not carry specific storage statements. In the EU, if there is evidencethat batches of the stored product as packed for sale are stable at temperatures up to 30" the product need bear no special temperature storage instructions. Within the US all drug products usually require statements on storage conditions. In Table 78 the storage instructio~s tobe used in the EU are listed. It means long-term testing, acc means accelerated testing. In principle, medicinalproducts should be packaged in containers that ensure stability and protect from deterioration. A label statement should not be used to compensate for inadequate or inferior packaging.Nevertheless, the statements in Table 79 maybeused to emphasise the need for storage precautions to the patient. All the results, the information on the tested batches, the applied analytical procedures, and the derived stability information, are compiledin the Stability report of the investigated drug product.

Storage Instructions to Be Used in the EU Testing co~ditionswhere stability has been shown 20°C f2"C/60% r.h. f 5% (It) 40°C f2"C/75%r.h. f5% (ace)

Required label 3

No labellingtobe

Do notrefrigerateorfreeze

Do notstoreabove

Do notrefrigerateorfreeze

used

2 5 ° C f 2 0 C / 6 0 % r.h.f5% (It) 30°C f2"C/60% r.h. f 5% (ace)

30°C

25°C 3r:2"C/60Y0r.h. f 5% (It)

5°C 3r:3°C (It)

--b 2"C--8"C Store at

zero Below

~dditional label,a where relevant

Do notstoreabove 25°C

Do notrefrigerateor

freeze

Dofreeze not

Store freezerb in a

a Depending on the dosage form and the propertiesof the product, there may bea risk of deterioration due to physical changes if subjected to low temperatures. Low temperatures may also have an effect on the packaging in certain cases. An additional labeling statement may be necessary to take account of this possibility. At a justified temperature.

General Storage Statements in the EU Additional labelling statementsa the package Comment

Storage problem depending on 1 2 3 4 a

Sensitivity to moisture Sensitivity to moisture Sensitivity to light Sensitivity to light

Keep the container tightly closed Store in the original package Store in the original container eep container in the outer carton

E.g., plastic bottles E.g., blisters

A rationale for the labeling statement should be given in the package leaflet.

According to the two different stability protocols for climatic zone IT and the climatic zones 111 and IV, two separate stability reports are written. untries of climatic zone 11with the EU, Japan, and the port for climatic zones I11 and IV. Thisreport contains as supportive data. For the climatic zones111and IV,different packaging materials and different specifications may be required. To illustrate the amount of information that will be available by applying the strategic planning, an example is given. The stability inforrnation derived for a registration application is basedon the results of Table 80.

Stability Information Batches

life

Three strengths of laboratory batches Three strengths of clinical batches phase I1 One batch phase I11 One registration batch Three registration batches

Investigations shelf

Derived

Stress and long-term Stress and long-term

6 months phase I 18 months phase I1

Stress and long-term Stress Accelerated and long-term

24 months phase I11 24 months 24 months

Thereby it can be assured that the patient after marketing authorization gets the same quality as the patient during the clinical trial investigations. The stability report for the drug product is structured and written in the same format as those during development. Stability information In the dossier for registration applications are included: rimary and supportive stability reports with the derived stability information. Testing specifications for release and stability testing. Validation report for the analytical procedures.

Required Capacity Required capacity and for a single analysis Dosage documentation and form

Required capacity for a time point accelerate^ long-term testing and on-going climatic zone I1

17 h P 0.45 weeks h P 0.32 weeks 18 8 h P 0.21 weeks

Solid Semisolid 12 Liquid

worldwide -

20h 0.53 weeks 32 h P 0.48 weeks 27 12h P 0.32 weeks 18

h h h

P 0.72 weeks P 0.48 -

Testing Analysis specifications including full

(weeks) validation Dosage fom

zone batches (weeks)

Solid Semisolid 3

Liquid

of test samples

No. ofclimaticclimatic I1

4.3 13 3 3.6 1 3 2.4 3

2.7 8.1 2.4 7.2 1.6 5

l1 7.2

weeks

_______

Total required capacity

worldwide zone

P 0.85 weeks

worldwide

7 12 7 l1 5 8

8 8 15 6 10

The required capacity is derived for 1 dosage and 1 packaging material.

onfir firm at ion and extension of the anticipated retest period, shelf life. onitoring of the stability characteristics for as long as the anticipated retest period/shelf life and the storage period are not yet identical, the data being generated by those studies already initiated or referred to in the registration application. Stability testing of three production batches ~anufacturedfor marketing after marketing authorization, if registration batches were not full scale.

The ongoing stability testing is a continuation of accelerated and long-term testing; it concludes the development of a drug substance or drug product.

. Selection of batches and samples Representati~ebatches of step 4. epresentative production batches manufactured for marketing. st criteria, analytical procedure, specifications as for step 4. ability test protocols (Tables 82, 83). rug substance and drug product

Climatic zone 11 (months) Batches conditions

Packaging material

Storage Storage period,

3 registration Simulating 25"C/60% batches proposed bulk storage container (drug substance)

testing frequency Testing specification 18, 24,36,48,60

r.h.

3 productionFinal packaging2SoC/60% r.h. batches formaterial 4OoC/75% r.h.* marketing (drug product)

0, 3,6, 9, 12,18,24,36,48,60 3, 6

corresponding to step 4

corresponding to step

*Drug Product only

Climatic Zones 111 and IV Packaging (months) Batches conditions Storagematerial 3 registration Simulating 3OoC/70% r.h. batches proposed bulk storage container (drug substance) 3 production

batches

Final packaging 3OoC/70% r.h. 4OoC/75Yor.h. material for marketing (drug product)

period, Storage

frequency testing 18,24,36,48,60

0, 3, 6, 9, 12, 18, 24, 36, 48, 60 3, 6

Testing specification corresponding to step 4

corresponding to step 4

aDrug product only.

Number of batches: 3 pilot plant from step 4; 3 production batches. aterial: drug substance: simulating proposed bulk storage cong product: final packaging material for m a r ~ e t i n ~ . e same methods and procedures are applied as for the registration batches. It may be necessary to derive post approval specifications from the data. The information on the robustness willbe extended for and the drug product by the following factors: atch size of starting materials Site of manufacture, chemical production Synthetic route, method of manufacture and procedure: final process Scale of manufacture: production scale Drug product: Different batches drug substance from final process and prod~ction scale Site of manufacture: pharmaceutical production Manufacturing process: final manufacturing process Scale of manufacture: production scale Stability reports with the extended stability information are written on demand, e.g.,

Yearly updating of the stability report At the end of the anticipated shelf life/retest period Stability information: confirmation and extension of the shelf lifelretest period and storage instructions if necessary.

7.1 . l . Objective Monitoring of the stability of the continuous production on firm at ion of the derived stability information 7.1.2. Application of the 6asic Principles Selection of batches and samples: representative production batches. Test criteria: selection according to results of the steps 4 and 5. Analytical procedures, specifications as for steps 4 and 5. Storage conditions, testing frequency, storage period, as in Table 84. Number of batches Since the first yearof production is coveredby batches for the ongoing stability testing, the follow-up stability testing starts in the second year. drug substance: 1 batch every second year drug product: l batch per year. If several strengths are marketed, only the most sensitive is investigatedor matrixing design (l/3 design) is applied. It is advisable to startwith the storage of the batch for a drug product in the same months every year. Packaging material: commercial.Ifseveral are marketed, only the most sensitive is investigated, or matrixing design (l/3 design) is applied. Evaluation: the results are compared with those of the steps 4, 5 and summarized in a yearly quality report. ~tabilityinformation: confirmation of the stability information, shelf lives, retest periods. The follow-up stability testing is not regulated by a stability guideline. It is a GMP measure to guarantee the quality. Storage Conditions, Frequency, and Storage Period Testing Storage conditions Climatic zone

1-11 60 111-1v 60

frequency, Testing

("C)

(% r,h+)

25 48 40

36 75

0 24

3048 40

7036 75

0 24

storage period (months) 12

6 12 6

Stability testing for variations and changes to a marketing authorization, post approval changes. The scope and design of the stability studies for variations and changes are based on the knowledge and experience acquired on active substances and drug products. rug substance Variations and changes of the synthetic route or manufacturing process of the drug substance may influence impurity profile, mainly by-products, intermediates, degradation products particle size and particle size distribution surface area polymorphism The following information is available: Justification of specifications Justification wasgiven for each analytical procedure and each acceptance criterion. The justifications referred to relevant development data test data of batches used in toxicology and clinical stddies results from stress testing results from accelerated and long-term testing results from production batches The batches for justification have been manufactured according to different starting materials different synthetic route different method of manufacture and procedure different site of manufacture different scale of manufacture Stability information The stability infomation is based on the following data and derived stability information: The results of stress testing, the stability profile with derived preliminary retest period The results of confirmation investigations with preliminary confirmed preliminary retest period The primary data with accelerated and long term testing up to registration application with derived preliminary retest period The ongoing stability data of the 3 registration batches, long-term testing with derived retest period of 60 months The ongoing stability data of the 3 post approval production batches, accelerated and long-term testing with confirmed retest period up to 60 months. Robustness of manufacturing process The batches investigated in stability testing were manufactured according to the following variations and changes: Different batches withdifferent batches and batch size of starting materials

Different synthetic routes that changed during development Different method of manufacture and procedure with different types and sizes of equipment Different site of manufacture, development laboratory, pilot plant, chemical production with different types and sizes of equipment Different scale of manufacture, laboratory, pilot plant, production scale On the basis of this detailed information it is usually possible to evaluate the influence of the variations and changes on the stability of the drug product. This is valid all the more for stable drug substances (stay within initial specifications 6 months 40°C/75% r.h. and 2 2 years 25"C/60% r.h.). There may be an influence on the following test criteria which have to be investigated carefully after production at release: particle size, particle size distribution, surface area, polymorphy, purity, dissolution rate. To be on the safe side the general procedure is proposed as in Table 85.

Drug Substance

Variation, immediate investigation results inchange Equipment ~anufacturing site Process Route of synthesis

Possible influence on influence quality, Possiblerelease specifications, Assessment Particle size Surface area Polymorphy Purity Dissolution rate

on stability, period retest Stability investigation

Quality none unchanged M1 data within specifications none none

2 batches

4O0C175% 3 months

rug products Basically the same amount of information is available as for the drug substance. Justification of Specifications Justification was givenfor the selection of the test criteria, each analytical procedure and each acceptance criterion. The justifications referred to Relevant development data Test data of batches used in toxicology and clinical studies Results from stress and conformation studies esults from accelerated and long-term testing esults from production batches

Stability information he stability information is basedon the following data and derived stability information: The results of stress and confirmation studies during the clinical development: different strengths and composition phase I derived: 3 or 6 months predicted and confirmed different strengths phase I1 derived 12 or 18 months predicted and confirmed final formulation phase I11 including registration batch derived: 24 or 36 months predicted and confirmed; 24 months predicted for registration batches The primary data of the registration batches, accelerated and long-term testing derived: 24 months preliminary shelf life confirmed The ongoing stability data of the 3 registration batches long-term testing derived: self life up to 60 months The ongoing stability data of the 3 post-approval production batches accelerated and long-term testing derived:confirmedshelflife up to 60 months Robustness of formulation The batches investigated in stability testing had been manufactured according to the following “variations and changes”: Different batches of drug substance and excipients including drug substance from production Different strengths Different compositions ifferent packaging material, type and size Different manufacturing process with different types and sizes of equipment Different sites of manufacture: development laboratory, manufacturing site of clinical supplies, pilot plant, pharmaceutical production with different types and sizes of equipment ifferent scale and batch site of manufacture laboratory, pilot plant, production On the basis of this detailed information it is very often possible to evaluate the influence of the variations and changes on the stability. In the individual case it depends on the dosage form and the type of formulation. The STJPACand C P guidelines and requirements do not refer to this broad base of information. By applying the strategic planning; scientifically based information can be provided in most cases. This information provides a higher degreeof certainty than formal stability testing of 1-3 batches stored at 4OoC/75% 3-6 months.

It is always necessaryto investigate these batches carefullyafter production to see whether all data are well within release specifications. It may be also necessary to perform special investigations. Nevertheless the following general procedure is proposed to be on the safe site:

Solid Dosage Forms on Variation, change in Equip~ent

Possible influence on influence quality, Possiblerelease specifications, immediate investigation

Assessment of results

Appearance unchanged Quality Content uniformityAll rate Dissolution

data withinrelease specifications

life

stability, shelf specifications. Stability investigation 2 40*C/75% batches or 3OoC/70%up to 3 months none

~anufacturing site Process

2 batches 40°C/75% up to 6 months

Excipient Qualitative Quantitative

7 SemisolidDosage Forms

Variation, change in

Possible influence on influence quality, Possiblerelease specifications: immediate investigation

Appearance Equipment Homogeneity Content Uniformity ~anufacturing within container Chemical site Stability Preservation Process Excipient: Qualitative Quantitative

life Assessment of results Quality unchanged All data within release specifications

on stability, shelf specifications. Stability investigation

2 batches 4OoC/75% (3O0C/7o%) up to 6 months

Liquid Dosage Forms Possible influence on quality, release specifications: immediate investigation

Variation, change in

Possible influence on stability shelf life specifications. Assessment of results

Stability investigation

unchanged Quality Appearance none All data within PH release Chemical Stability specifications ~anufacturing Preservation site

Equipment

2 batches

4OoC/75%

up to 6 months Process Excipient: Qualitative Quantitative

Change in Packaging Material

Variation, change in Immediate packaging Same material different size

Different material

Possible influence on on influence quality, Possible release specifications, immediate investigation

Assessment of resufts

life stability, shelf specifications. Stability investigation

Solid dosage forms Quality unchanged All data within release Tightness of specifications Container Semisolid dosage forms Homogeneity Content ~ n i f o r ~ i t y Liquid dosage forms Chemical Stability

none

Solid dosage forms Permeability 02, H20, Light

none

Equal or less permeable Higher permeability

Test procedure

Semisolid, liquid d.$ Permeability 02, H20,Light Interaction

Equal or less permeable, no interaction

Specificity Sensitivity Validation

Corresponding validation data equal or better

none none

2 batches

4O"C/75%

up to 3 months 2 batches 4OoC/75% or 30°C /70% up to 3 months none

After thesegeneral follows.

statements the application of the basicprinciples

7.2.1. Application of the Basic Principles

Selection of batches and samples: drug substance:pilot plant; drug product: the manufacturing process to be used should meaningfully simulate that which should be applied to large scale batches for marketing; the quality must meet all release specifications. Test criteria: corresponding to original formulation. Analytical procedures: corresponding to original f o ~ u l a t i o n . ~peci~cation: corresponding to original formulation. For storage conditions, testing frequency, and storage period, see Table 90, umber of batches: 2. ackaging material: commercial unless packagingmaterial has been changed. valuation: the results are compared with the corresponding data of the original formulation. If no new data (2 3 years) are available, 1 batch of the original formulation is investigated together with the changed formulation. Stability information: if the variation causes no infiuence on the quality, all data are within specifications;the stability information of the original formulation is still valid. The stability is persuedby the corresponding follow-up stability program, but in the first year of production two representative batches are put on stability. urthermore the testing frequency at 40°C is 3 and 6 months. If the changed formulation is less stable the shelf life has to be shortened accordingly. Conditions, Frequency, and Storage Testing frequency, storage period (months)

Storage condition (“C) (% r.h.) 40 30a 25

75 70a 60

1 **

(6) (6) 6

up to shelf life a

Only if at 40°C significant changes are expected.

1. The Tripartite Guideline on Stability testing of new drug substances and products endorsed, Q1A ~hotostabilitytesting, step 4, Q1B Test on validation of analytical procedures, step 4, Q2A Extension of the text on validation of analytical procedures, step 4, Q2 I~puritiesin new drug substances, step 4, Q3A I~puritiesin new drug products, step 4,

3. 4. 5.

6. 7. 8.

Residual solvents, step 4, Q3C Specifications, step 4, Q6A W Grimm. The extensionof the ICH Tripartite Guideline, SecondInternational Meeting of the Southern African Pha~aceuticalRegulatory Affairs Association, 15-17 March 1995, Pretoria. PMA’s Joint-PDS Stability Committee “Stability Concepts,”Pharmaceutical Technology (1 984). Krummen. Stability testing in the EC, Japan and the USA. Wissenschaftliche Verlagsgesellschaft, 1993, p. 17. W Grimm. Stability testing of clinicalsamples. Drug Development Ind. Pharm. 22235 1-871, 1996. WHO Expert Commitee on Specifications for Pharmaceutical Preparations34. Report, WHO, Geneva 1996. W Grimm. Extension of the ICH Tripartite Guideline for Stability Testingof New Drug Substances and Products to countries of climatic zones 111and TV. Drug Development Ind. Pharm. 24(4):319-331,1998. W Grimm. Drug Develop. Ind. Pharmacy 19:2795-2830, 1993. Revised version of Annex 14,Manufactur~of investigational medicinalproducts, of the

This Page Intentionally Left Blank

~onsultant,Beeston, ~ o t t i n g h a England ~,

484

1. Introduction ackaging Protection Functions 2.1.Physical hazards 2.4.Chemical

hazards

485 485 486 49 1 492 492

covering blown and tubular containers layer materials and coatings er based materials and elastomers 4.6. Standards for sterile products

493 494 494 495 496 496 496

5. Testing the Finished Product in its Pack 5.1.Leakage detection methods 5.2. Evaluating rubber stopper assembly

497 497 499

6, Legislation and Guidelines Relating to Packaging 6.1.Typicalpackaging information normally required velopment pharmaceutics kaging materials (primary or intermediate packaging)

499 50 l 501 501 50 1 502 502

6.6. Control options 7. General Observations on PackagingSelection

503

8. Likely Impact of the IGH Guidelines on Packaging

isk Versus Intensity of Testing 10. ~ntensityof Testing List oints Requirin~Special Consideration and for Guidelines 13. Conclusions eferences

50 508 508 510 512 512 512

~ l t h o u no ~ hpha~aceuticalproduct can be licensedwithout a tested and approved pack, the complex functions of packaging are fre~uentlynot fully recognized. Since most in industry are likely to have at least an indirect relationship with a packed product, all should be aware of what packaging basically sets out to achieve. this is initiated by some form of shock treatment, i.e., have you ever seen which is not in a pack-yes-as a puddle on the ground-thereby indicating that all liquids need to be contained in an effective way. aving indicated t need for a pack, let us now provide a clear basic definitionof packaging is the economical means by which a product is rotected resented rovidedwith

identification, information, ~ontainment,co~venience,and

USE with due attention to ultimate DI ~ ~ (i.e.,Sthe current ~ L environmental issues). This invariably equates with the ~ r o ~ ushelf c t life. ince the final shelf life, which is declared depe t on the pack employed, all ent activities must bechecked against the (~nternationalGonference nization) Guidelines (1). This chapter therefore hi~hlightsthese, in~icating where there are possible deficiencies in the guidelines, which haveto be covered as part of the pack approval program. Topics are also stressed where the g~idelines make assumption^'^ that certain work, critical to general product and drug subough all the factors identified inthe definition of packaging contrib~te to the overall function and performance of the pack, the word protection tends to be most closely aligned with shelf-life. Protection is related to

hazards and is therefore a major function, which is described in detail first.

Packs meet physical hazards both when theyare static (stored, displayed, etc.)and in motion (during distribution and any stages of handling). The main physical hazards can be quantified under the headings of 1, Impact /shock (dynamic hazards) 2. Compression (static hazards)

2.1.1.

irnpact/Shock

Impact involvesforcesimposed on the product and pack by acceleration/ deceleration. These forces may cause breakage, fractures, distortion, dents, etc. and may be transmitted via the pack to the product, i.e., damaged tablets, broken glass bottles, damage to closure systems, etc. 2.1.2. ~ o r n p r e s s i o n

This arises from stacking loads both during warehousing and in combination with vibration during transportation. Whether thiscausesproblemsdepends on the way the stack is constructed, the stacking period time, the nature of the bearing surface on which the stack is built (i.e., may relate to the type of pallet and its “footprint”), the environmental conditions around the stack, and the nature of the goods underthe compressive force. The strength of the stack may also be modified by dynamic hazards such as vibration and impact. 2.1.3.

Vibration

Stress causedby vibration depends on amplitude and frequency and may arise from vehicle engines, road surfaces, etc. It is usually measured in hertz-i.e., one cycle ajor damage usually arises between 3 and 30 hertz. ~ i b r a t i o ncan ckproblems-rubof surface, decorated or undecorated? possible loosening (rather than tightening) of screwcaps, increased electrostatic o etc. and product problems-particle size separation, product powderi 2.1.4.

Puncture, TearandSnagging

These problems may occurexternally or internally to the pack. The risks from the above can be quantified by actual travel and warehousing tests, simulated (laboratory) type tests, coupled to the useof instruments in packs during tests. The latest instruments, called data loggers or acquisition units, can quantify various actions as functions of time, i.e.,temperature, humidity, impacts (drop heights), and vibration levels. Such information can then be utilized to assist simulated testing. ‘er,the above “hazards” can occur in isolation, but more often combination. Typical stability testing involves basicallystatic

climatic tests, which do not include anyof the physical challengesjust mentioned. It is,however,essential that these be evaluated, usually as part of a package investigational program. This may also include storage periods under a range of climatic conditions followed by analysis to check that no critical changes have occurred either to the product or to its pack. The author recalls one experience where a fine powderwas adsorbed onto a larger crystalline carrier. This was packedinto a sachet and immediately analyzed. 100% of active was identified. The sachets were then transported 80 miles to the stability testing area, where the product, when dissolved in water (as an oral drink), returned only 95% active present. The other 5% was found adsorbed to the walls of the sachet. Initially this was counteracted by a 5% overage followed by ultimate reformulation of the product. A non-travel tested product did not initiallyshow this activeloss.allelproblems can arise with closures,whichmaychangeinsealefficiency ac ing to the influences of top pressure(compression), vibration, impact, etc., c d to expansion or contraction due to temperature/humidity variations.

The previously mentioned physical hazards can be further influenced by the prevailing climatic conditions surrounding the pack. For example, an atmosphere of high humidity can reduce the stacking (and compression strength) of corrugated boxes and humidity increases, by well over 50%. imensional changes,due to temperat~re can also arise, especiallywith r based materials, where expansion issignificantly greater than for metals. ial permeabilitychanges also vary according to circumstances and hence may not follow the various “orders” of reaction or directly relate to anArrhenius plot. The climatic hazards, which may influence product stability therefore, have to bediscussed in terms of pack properties and characteristics. 2.2.1. ~ o j s t u r e - A s L i q u i d W a t e r o r ~ e l a t i v e H u m i d i t y (

Although glass and metal (when freefrom perforations) are totally impermea~le,all plastics are to some degree permeable. This may result in either moisture loss or moisture gain, depending on the nature of the product, the characteristics and the thickness of the polymer, and the “gradient” between the inside and outside ntaining certain “salts,” when packed in a plastic k, An aqueous produ may show no moisture loss if the internal vapor stored at 40°C 75 ressureis at equilibrium with the external atmosphere. Storing at uld show a significant moisture loss, as there is a very positive “gradient” n the two atmospheres. here products take up moisture, the greatest moisture gain will arise when external temperature and humidity ishigh, i.e., 40°C 75% n the biscuit (US,cookie) industry, where a product such a up moisture, becomes soggy, and looses its “break” ‘spness, an acceleration factor of 2 x is often used for each rise of 10°C and 1 ,provided thesefactors ompare 40°C ’75% do notexceed an acceleration factor of 8. If this was biscuit storage condition of 20°C 4 ,it would give a (false) of 2 x 2 (temperature) and 2 x 2 x ,giving a factor of 32,

which may give some indication as to the severeness of such a condition based on simple moisture gain and not involving a chemical reaction. Permeation also depends on the surface area exposed, the material thickness, and how the permeating material is removedfrom the far “environment” so that the gradient conditions are retained. Bottle-to-bottle contact (especiallywith rectangular, square, flat surfaces), or shelf-to-bottle contact (particularly if the shelf has no perforations or is enclosed) will influence the rate of permeation-since contact with the permeating layer is impeded or reduced. This thought can be extended further, as most packs have a series of enclosures (the secondary packaging) all of which slow down any loss or gain significantly, i.e.,

1. A blisterpack listers in a carton 3. Cartons in a display outer 4. Display outers in shipping outer 5. Shipping outers on a pallet stretch or shrink wrap on a pallet. This results in a steady improvement in the total barrier for the product. Some of the above can be additionally overwrapped, thus building up the barrier properties. The author once didan experiment in which a blister (exposed“naked” at 37°C 90% RH) had a shelf life of three weeks, indicating that a blister was an unsuitable pack. The same pack, stored somewhat likethe above but with a carton and display outer overwrap, was placed inthe company’s warehouse wherethe product was still in “specification” after 6 years (3 weeks to 6 years represents an “acceleration factor” of over 100). 40°C with extremes of humidity, 75% RH to 15% RH (hot moist to hot dry conditions), provides a high challenge to packaging materials, components, and finished packs. Cellulose based materials are particularly susceptible to moisture/ temperature changes, as most properties are related to moisture content (and grain direction). It is therefore essential to “condition” such materials to 23°C 50% RH if comparative testing is to be meaningful. Conditioning maybeadvisable for other materials (e.g., plastics), if properties can be changed by external factors, 20°C 65% RH being used for polymers. Like products, moisture is a critical challenge to many packaging materials-as even in the form of condensation (on glass, plastic, metal, foils, etc.) problems can arise (i.e., encourage bioburden, alter surfaces,set up erosion, corrosion, etc.). Some of these aspects will be discussed further under packaging materials and general combination effects. Moisture (and gases) can be soluble (S) in and diffuse (D) through a polymer, hence permeation (P) is related to both these factors (P = SD). Moisture can also act as a carrier for other permeants, i.e., foreign or actual product flavors and aromas, organic and inorganic gases, depending on their solubility levels. 2.2.2. Gases

Cases, inorganic or organic, can pass through certain materials, The resulting interactions may relate to chemical changes, pH, flavor and aroma, color change, etc., any of which could beundesirable. Although oxygen isoften the more critical, other gases or vapors that may be present, must not be excluded from consideration

without some prior assessment. For example, the permeability ofthe common gases through plastic generally followsthe ratio 1 :4 :20 for nitrogen, oxygen, and carbon dioxide. This usually means that anaqueous solution of a nonbuffered product will slowly move from say pH 7.0 to around pH 4.5 due to the presence of carbonic acid formed by CO;! permeation, if it is stored in a plastic pack. In the case of LDPE (low density polyethylene),this may only take 3-6 months, whereas it occurs more slowly with a PET (polyester) bottle. If the pH then drops further, to pH 4.3 or lower, this is usually an indication of polymer degradation involving the shorter chains. Many polymers can suffer from oxidative changes and hence may contain antioxidants, stabilizers, whichreducesucheffects,especiallywhencoupled to thermal challenges. The EP (European Pharmacopoe~a) refers to grades of low-density polyethylenes which are free from antioxidants, but these need extra assess~entrelated to potential degradation. Note certain polythene films if used out of doors may show rapid degradation due to combined effectsof light, oxygen, and other gases, heat, moisture, etc., but if kept indoors, little deterioration occurs over a 5-10 year period, This also puts emphasis on the possible needfor ‘‘surfa~e’~ analysis, since gasescan also alter the surface properties of any material, e,g., glass, metal, plastic. The author now argues that all materials are different at the surface when this is compared with the main “body” of the material. The surface is also the main interface between packaging material and product, where any interaction/e~changestarts. Thus gases,whichmight not change the product, can have a greater influence on the properties of the pack. ( aveyounoticed how manynatural polymers, when placedout of doors, can rapidly deteriorate, i.e., film, strapping and even rigid containers?) 2.2.3. Light

The pack and its packaging components may involve twoquestions with referenceto light: 1. Do they excludeor reduce lightfrom reaching and possibly influencingthe product? For example, amber glass filters out a large proportion of UV light. 2. Does the pack absorb any particular range of light rays, and does this change the properties of the packaging materials involved?

For example, amber glass absorbs IR (infrared) rays, which can heat up the product and pack: colored surfaces may discolor, i.e., darken or fade, depending on the intensity of the light. Again, light may operate in conjunction with other factors to cause various forms of degradation or change, which again start with surfaces, e.g., rubber products are influenced by heat, light, and oxygen and hence may react differently to needle penetration, fragmentation, coring, etc. after “aging.” Retests for these are essential during prolonged periodsof any product storage-and further changes may also arise from product contact. Tests for light exposure may involvesimulated tests, or actual conditions, i.e., behind glass, north or south facing windows, or total exposure to all the elements (wind, rain, sun, general atmosphere, temperature, etc.). Controls may consist of total darkness (a condition found in many climatic rooms and cabinets) or an area of the product-pack masked by a 100% light barrier material. It is important to

l~ation,Sta~iiity,and Shelf-Life

understand whether this acts as a “white body’?or a “black body”: one reflects the heat/light rays, whilst the other is absorptive and hence may cause heat related differences. Materials themselves will also have heat reflective or heat absorptive properties, etc. 2.2.4. Pressure

Atmospheric pressure differentials can occur both internal and external to a pack. These can arise due to a number of reasons: a. Natural fluctuations for any specific height above or below sea level, e.g., changes in altitude due to geographical location. b. Transport in an unpressurized aircraft flying at 10,000 feet-creates a negative pressure of approximately $ atmosphere. c. Transport in a pressurized aircraft. Such aircraft are pressurized to an equivalent of 8000 feet, whichis 3.8 psi (pound per square inch) or roughly 4 atmosphere. (One atmosphere is14.7psi.)Thismeans that items are subjected to 4 atmosphere on ascent and 1/ 4 atmosphere on descent to sea level. d. Hot fill plus effective closuring and cold fill plus effective closuring can create negative (vacuum) and positive pressures within the pack, when the pack is subsequently stored under normal conditions, e, Moist heat sterilization of filled packs, where the pack may vent and then reseal or simply expand and then not fully contract-as this may create a negative pressure within the pack. With certain flexible polymers this may give rise to pack “dimpling” or partial collapse due to negative pressure. (This can be partly controlled by an over pressure or a balanced pressure autoclave.) Negative external pressures cause flexible and pliable packs to extend, and positive external pressures cause such packs to be under pressure or compression. Under these conditions packs that are not effectively sealed will “breathe,” i.e., let ~‘air”/moistureetc. in and out, according to the changing conditions.

2.2.5. OtherAirborneConta~ination

This can involve particulates and bioburden. Particulates can arise from various sources and also be a carrier for bioburden. Particulates can come from the packaging material itself (i.e., fibers from board and paper, wood based materials, pallets, glass particles from cutting glass,etc.) or a secondarysource (hairs, fragments of skin from human beings, etc.). The generation of particles from basic materials can occur in many ways, i.e., when they are subjected to vibration/ abrasion, cutting/ guillotining/punching out, breakagelimpact /chipping, etc., hence the inference that the least handling and processing will probably lead to the lowest particulate levels. In converse, greater handling will encourage particulates, and this is partly counterbalanced by more use of“vacuum extraction’’ and environmentsusing “filteredair.” Study of air filters is then one way of providing better identification of the particles removed. On production lines, such operations as unscrambling, use of vibratory bowls, abrasion from moving conveyors, etc.can all be a source of particles, much of which “may7’ be removable by effective “vacuum extraction.” The materials on which (wooden pallets) and in which packaging components are delivered (fiberboard) and stored tend to be a serious potential source of extraneous particles. These materials may require ongoing inspection, cleaningto remove surface dirt, or actual banning

ean

from certain types (of classified) production areas. Pallet control can be an important part of this operation, both for incoming supplies and outgoing finished stock-as too frequently this can be overlooked. Isolating filling and closuring from other activities can also improve cleanliness. Particulate control, particle identification, and particle sourcing, together with the same for bioburden, are all part of cGMP (current good manufacturing practice). Finding an excess of particles (or bioburden) at any stage of development through clinical trial supplies, formal stability tests, or final production, is a clear indication of poor cCMP. However, finding a product with an acceptable level of particles does not necessarily mean that this “state” cannot change. Particles can reduce (possibly by adsorption) or increase during both storage and transportation, for a number of reasons. Particlescan arise from packaging materials that have a surface growth, as crystals, as a bloom, simplemigration, etc. This has been knownto occur with rubber based materials, plastics generally, and even glassby a process of surface erosion. These will be further discussed under properties of packaging materials, e.g., surface activeconstituents in polymers that can be chemicallyor physically removed by surface abrasion. All of the above relates to general “cleanliness,” whichoften means producing materials “clean” and then handling them under controlled conditions, using specified procedures in order to optimize total cleanliness. The alternative is to produce effective cleaning processes on line, which may involve actual “washing” or some alternative means. For example, “air blowing” must use pressurized, dry, oil-free, particle-free air of a controlled temperature, coupled with vacuumextraction, using inverted containers (invariably inverted in the case of glass), if effective cleaning is to be achieved. Again, this air should be “filtered” and “captured” so that checks can be made for types of particulate and possible sources by a regular follow-up, i.e., it provides a continuous learning opportunity. “Cleanliness” has also to be balanced against general “hygiene,” especially in terms of training, clothing for operators, working practices, etc. 2.2.6.

PrintingandDecoration

If a packaging material undergoes a decoration or printing stage, this is often seen as a relatively “dirty” operation, hence contamination can arise from these processes. Contamination mayinvolve particulates (whichmay be ink particles), solvents (although there is now a tendency towards nonsolvent inks), and general migratory substances, which may be of polymeric origin. Abrasion of printed surfaces (rub) is a further source of particulates. Once rub starts, the particles produced can increase the possibility of abrasive effects. Checking hygiene and cleanliness istherefore part of all stages of product-pack evaluation and general good manufacturing practices. 2.2.7. Aging

Aging istraditionally accepted as change(s) against elapsed time, hence is based on more than one cause and effect. Like all of us, both products and packs, and the constituents from which theyare manufactured, can age. This is currently being recognized with packaging materials; many companies are now restricting ‘61ife” by a “reexamination date” or “reinspection date” until a factual shelf-life can befirmly established. As indicated earlier, any change can be related to either

the material surface or the main “body” or both. Knowledge of these aging effects is currently in its infancy, particularly as to whether they will have any influence on the product-pack shelf-life. Surface evaluation can include chemical analysis, appearance, and in many cases microscopical examination for cracks, pores, reticulation, smoothness, etc. and physical examination, e.g., coefficient of friction, etc. In the case of metals, where surface area can be increased by abrasive substances, corrosion maybe increased. Stress cracking of polymers is another factor that can beinfluenced by the nature of the surface.

These can include microbiological (bacteria, molds, yeasts, etc.) toxins and related substances produced by them, damage by animals, including vermin, birds, insects, reptiles, etc., and hazardous materials associated with them (e.g., excreta), and last but not least, human beings. Packaging aspects, related to the latter, may involve prevention or restriction of access (e.g., tamper resistance, tamper evidence, child resistance, i.e., security related) or the opposite where ease of access and/or reclosure is of importance (e.g., for the ever-growing elderly population which survive by taking regular medication). This must be supported by unambiguous, easy to understand, legible instructions, coupled to the avoidance of misuse or abuse. Since microbiological control may involve the use of preservatives, aseptic presterilized materials or terminal sterilization materials have to beselected to meetthese demands. The possible presenceof pyrogens, mainlyassociated with liposaccharides from dead gram-negative bacterial cell walls, also has to be considered. The use and pack function in the administration or drug delivery aid is an important evaluation stage, as any changes in product and pack must be assessed both separately and together, i.e., does a new pack or anaged pack effectively deliver a product at the beginning and the end of its shelf-life. Thisalso involves risks associated with anyform of contamination. However, assessment may not be this simple, if the component delivering the drug can be influenced by intermittent or infrequent use (e.g.,a pump based systemor anaerosol) or is used ina separate device (powder inhaler or a nebulizer unit) that may be subject to “cleaning” by the user. Such delivery systems may differ as follows: 1. The pack directly acts as an administration aid (drops, sprays, prefilled syringes, aerosols, pump systems). 2. Thepackis a “feeder unit” for a separate device-Glaxo-Wellcome Spindisk for Diskhaler. 3. There is a separate device into which product is transferred for administration purposes (nebulizers). 4. The deviceis manufactured containing a feeder unit (Glaxo-Wellcome Accuhaler).

Each of the above can lead to complex proceduresto check that the total performance of pack, product, and administration system function satisfactorily. The above form “gray areas’’ with true devices, and hence consideration of “device” guidelines may be essential in the evaluation of these systems.

~hemicalhazards initially relate to compatibility between the product and pack and any “exchange”that may occurfrom product to pack or pack to product. These may involve various chemical interactions or simple mi~ration/leachingwhere no initial chemical change occurs. These typesof situation may cause changesin appearance (e.g., corrosion of metals), aroma/flavor, microbiological integrity, preservative activity, pH, etc., as well as product potency, bioavailability, etc. Like products, packaging materials and components have levels of purity, impurity, and possible residues arising from the conversionprocess or method of manufacture (i.e., polymerization process in the case of plastics), plus any constituents added to modify or improve the material in some way (glass, metal, and plastics may allcontain such terials that are from sources of natural origin, i.e., ores, metals, to vary in the i~purities(types and levels) according to their ~eographicalsource. Although these factors may not be critical in their general use, they do tend to be partly overlooked and could be important when the end use is for large- or small-volume parenteral products, where the risk of the “user” and “company” may be the greatest. Guidelines. Normally any subProduct purity is an essential part of the I gradation product needs identistance present at a 0.1% level as an impurity o fication and characterization with reference to its safety. urity of packs has yet to receive similarattention, although work maybe carried out on “extractives” from a chemical and biological point of view (5).

Since some ofthe hazards that can affect pack do not form part of any formal stability program, it is often assumed that they are adequately covered in the research and product-pack developmentstages, by investigational/feasibility testing or evaluation. However,these assumptions covermanysimplebasic factors that may act in a variety of combinations. Hence it is relatively easy to overlook an important point or even points. It is therefore essential to have an 660verviewer” who understands all the aspects of development and research, whereby there is a total data philosophy covering all stages from discovery of a new chemical entity (NCE) to ultimate product withdrawal from the market. This total dataphilosophy provides for accumulation of information from all stages, i.e., drug discovery, initial screening and safety-testing,human volunteer studies, early clinical evaluations, formulation development, formal clinical studies, investigational tests between product and pack, formal stability tests, etc. These coverthe Phase T to IV stagesfrom which the drug substance and the product dosage form are finally registeredand licensed as a packed product for specific diseases,etc. This data should clearly indicate that the product is safe, effective,and of the correct quality, well before anyformal stability test is carried out. This latter stage should primarily establish what already has been proved in terms of the shelf-life and stability of the packed product. A formal stability test should not therefore contain any surprises, assuming that adequate prior testing has been carried out. Reducing the cost of complex stability programs by and bracketing also makes similar a s s u ~ ~ t i o n s .

elf-Life

l1 vary according to the product ca r) verterinary, etc., its area of use/ in the body, the potency of the drug substance, which in turn can be related to risks to the user/patient and the pharmaceutical company producing it (product liability). is hence possible to produce a table that reflects this risk (especially for ethic products) and the level or intensity of testing advisable to minimize the risks. Such a table invariably is headed by large-volume parenterals with IV solutions at the top, followed by small-volume parenterals, with solid dosage forms at the foot of the table, where risks are minimal. To datemost guidelinesare related to ethical products, and their generic counterparts, becausethese carry higherrisks than OTC type products, which are involved in self-medication. Since ethical products usually start with a research program involving drug discovery and development, leading to a new emical entity, this has to be fully characterized and an approved pack established. 0th drug substance and the product dosage form must therefore be effectivelyevaluated prior to entering a formal stability program. Although both programs involve a wide range of challenges, in the case of the product dosage form, this usually comes under the heading of preformulation studies. The important activity of selecting the correct pack extends beyondcompatibility between product and pack and covers functional aspectssuch as efficientwarehousing and distribution, good production line performance, highcompliance and convenience in use, etc., plus general aesthetics. These may involve of use decorative techniques, color, size, shape, instructions, legal requirements, particularly where essential to sale, including name, brand image, and overall recognition features to encourage good presentation. Such factors may also be required to assist general CO and help instilladditional confidence in the product for both the profession involved (doctors, dentists, nurses, pharmacists, etc.) and the final user/patient.

aterial and pack selection requiresa good basic knowledgeof all materials, a plus (QC plus covers any range of tests to fully quantify a previously little-known material or component) system that quantifies exactly what is being used in any test, and an effective schedule by which product packs are evaluated. An outline of the main packaging materials follows. In practice this knowledge should cover: 1. The raw basic materials 2. Any purification of these raw materials 3. Conversion processes used to create basic packaging materials, components, etc. 4. ow materials or components should be packed 5. eneral storage and distribution 6. elivery and unscrambling on production line 7. eneral storage and distribution 8. Delivery and unscrambling on production line 9. assembly methods into a finished product/pack 10. s of storage and distribution and relative packaging methods covering handling and use options

11. Disposal options for each;includingrecovery,recycling,reuse, awareness of energy and pollution levels 12. Supportive test methodology for all of the above

and

As basic knowledge buildsup with experience, confidencein the pack properties and characteristics should increase. Someof this pack detail is therefore outlined below.

Glass exists as Type I Type I1 Type I1 Type I11 Type IV

neutral or borosilicate glass treated-asType I but surface treated by sulfating or sulfurizing glass-sodaglass-surface treated by sulfating or sulfarizing or soda glass or NP (non parenteral in the USA)

Soda glass is mostly used for ointments, creams, and oral and topical products such as liquids, emulsions,and solid dosageforms generally. It has a typically alkaline surface and may react with alkaline substances and leach out alkaline earths, especially if subjected to multiple autoclaving. Type I or neutral glass is widely usedfor small-volume parenterals, where low leaching is essential.Treated Type I glass isalso found where an occasional container might have failed the Type I test for hydrolytic resistance. Type I1 glass was widely used for IV solutions but is unlikely to withstand multiple autoclaving, as the “ n e ~ t r a l ’surface ~ can be lost. (Type I1 is a soda glass treated by a sulfuring or sulfating process to give a thin layer of coating which reduces the alkalinity.) Alkaline substances such as sodium or potassium tartrate, citrate, or salicyclate can cause greater extraction or create glass “flakes,” especially where soda glass is used,and the pack is subjected to steam autoclaving. Possible extractives from glass are now receiving more attention.

Metal containers can be made from tinplate, aluminum, or alloys of aluminum and stainless steel for large vessels. Surface attack can usually be eliminated by the useof external enamels or internal lacquers. Both of these consist of polymers of a thermosetting or thermoplastic origin and henceneed to bechecked for extractives or possible interaction with the product, both short and long term. Possible corrosion effects also need to be assessed. Although aluminum collapsible tubes are still used, multilayerlaminates are being introduced for many ointments. Aluminum foil is widely used. Most foils are surface coated with a thin polymer layer (washor key coating) which isadded to assist the “key” of the printing inks or other coatings (heat seal layersand laminations) and to reduce surface scuff, Foils of around 17 pmand above are usually commercially freeof pinholes-which increase in number with the thinnest foils (down to 6pM). Moisture permeation through these can be negligible, provided the foil is laminated to a plastic ply. However,

thinner foils can stretch, perforate and suffer from sealing deficiencies that may allow moisture exchange (inor out), depending on the type of equipment employed. etal-lisatin and other coatings are replacing some typical foil uses.

Today, plastics tend to be preferred to metal or glass, providedan adequate shelf-life can be achieved. Basically all plasticsare to some degree permeableto moisture and gases and need special requirements if they are to exclude light. They are also less inert than glass and may sufferfrom migration, leaching, absorption, and adsorption under certain circumstances.Fewplasticsconsist of the pure polymer (the EP mentions LDPE as one such material), and hence they may contain residues from the process of polymerization, additives (substances deliberately added to change certain properties), processing aids (added to aid processing), or master batch constituents (e.g., wherea concentrate is added and “let down” to the required content, ing “full” knowledgeofthese factors is relae.g., titanium dioxide of 1--3%). tively noncritical for solid do packs but can be critical for large- and small-volume parenterals, eye preparations, etc. Among the additives, there are constituents that are “active” by being present at the surface of the polymer, e.g., antistatic agents, slip agents and antislip agents, etc. These can be removed by surface abrasion. Many “poly” bags have a surface layer of a lubricant, typically “stearamide” or a metallic stearate. The loss of certain preservatives by absorption (loss into) or adsorption (loss onto the surface), particularly with LDPE, is generally well recorded, i.e., phenol, chlorbutol, 2 phenyl ethanol are absorbed and thiomersal can be adsorbed. The latter also depends on the nature of the surface-the larger the surface area the greater the loss.Such a phenomenondoesoccurwith other polymers. Certain polymers, especially LDPE, may also suffer from ESC (environmental stress cracking), which may occur when a polymer is stressed in contact with a stress cracking agent. The stress may be in-built or applied, and typical stress cracking agents include wettingagents, detergents, and certain volatile oils. Exchange of constituents between product and polymer can occur either way-product constituent(s) into the polymer, and polymer constituent(s) into the product. This always requires a thorough check, as some exchange, howev minute, is almost inevitable. Asa broad guide “liketends to absorb like,” e.g., EVO (ethylene vinyl alcohol) readilyabsorbs water and to some degree alcohol. Plastics may undergo changes during conversion and general processing.e.g., certain plastics need a pretreatment stage to improve print key. This is usually a surface oxidative process achieved by gas flaming (containers) or corona (high voltage) treatment (films). Both processes modify the surface properties, which may reflect on how any exchange occurs. The treated surface usually losesits properties with time. Sterilization processes tend to need evengreater investigation and control as all either tend to modify the polymer in some way or leave possible residues (i.e.,as in ethylene oxide sterilization). Gamma and beta irradiation can cause changes in the polymer chain (i.e., cross-linking, chain scission) and the constituents that they contain (e.g., antioxidants). Intense UV light (as usedfor surface cleaning) can also cause problems. It should be stressed that plastics provide an excellent service to pharmaceutical packaging, provided proper attention is paid to the above factors

and many others. A book, however, could be written on this to (polyethylene naphthalate), (liquid crystal polymers), which are conjunction withpolene or polyester, COGS ous polymers produced by metallacene catalysts, cyclopentanebasedpolyolefin from aikyo Seiki), and various coated versions of e~isting,widely used, polymers.

t growth in multilayer materials9 new gs, etc.), and new coating techniques. stic film, and coatings. polymer, hence previous comments er, there are welloverolymers that mightbeemployed, each subject therefore can becomevery e by lamination, but only polymers can be employed where coextrusion is involved.

ric based materials are generally more difficult to clear than tly because theyare involved in sterile products and have seleclability after needle penetration) that have not been achievable are widelyused as plungers in syringes and for closures in requires specific properties, i.e., ease of movement, ease of insertion^ freedom from particulates, low in coring and fragmentation, easy needle penetration, effective resealability, etc. Some of these r~quirementsare in conflict with other factors, so a compromise has to be achieved. Synthetic based rubbers (nowmainly butyl, chloro, and bromobutyl) offer good inertness, good barrier properties to gas and moisture, compatibility withpreservatives,goodage resistan~e,etc., but have poorer resealability, coring and fragmentation when compared with the earlier natural baseder materials. Adding a facing of P (polyt~trafluoroethyle ove inertness but puts extra demands on closure efficiency. aterials (natural or synthetic) contain nts all of which need screeningin terms of possible extractives ased materials find uses other than for sterile products. Sterile ear” in terms of all the materials used for a pharmaceutical product, includingchanges that may arise due to agi EU or IS0 (Inte~national the tests are nowcovered bynew ganization) standards.

tandards for sterile products include limits on bioburden prior to sterilization, tests for sterility, and a range of tests on basic materials that includepackaging ver from National stanc o ~ p o ~ e n tIn s . certain areas, IS0 standards are tak standards include those dards and s om pen dial standards. The more recent in the table.

IS08362 Injection containers for injectibles and accessories IS08871 Elastomeric parts rdness, fragmentation7self sealability, needlepen et ration^ seal S08871 Covers extractive tests, includingchemical and biol UV, reducing substances, a ~ m o n i anonvolatiles, , turbidity, volatile sulphur, etc. Validation is also of particular importance to sterile products and to all processes of sterilization, covering temperature recordings, bioindicators, dosimeters7 essure sensors, chemical methods of monitoring, and special asuring filter integrity, etc. In the validation of sterilizers the now accepted stages installation, operation, and performance of validation must be in qualifications, etc., i.e. *

on the pack and the scale involving accelllenges both on the

to perform. Tests are essential for all the features not covered by

therefore follows.

Leakage (egress and ingress) can be related to liquids, solids, or gases including bioburden and the methods associated with their detection. The methods for detection are becoming increasingly complex and more sophisticated with a trend, certainly for production materials, towards nondestructive testing. The listbelow coversawide range of tests, which are not listedin any particular order of servation of visual defects, e.g., pinholes, capillaries, etc., using human or automated inspection systems. Vision systems couldbe included under this heading-see also 12 should also be included. eight change-loss or gain against time under specifically defined conditions. A slow but reliable method. 3. Pressure-vacuumchanges by the application of pressure or vacuum under defined conditions (including fluctuating conditions), e.g., pressure change g various liquids including water, with and ase of gases-e.g., air-bubble type tests.

4. Caseous detection tests, i.e., helium(sniffing),e.g.,usingmass spectrometry leakage down Pa-m2/s can be detected. Oxygen,e.g., uses a dry pitrogen stream whereby oxygen is detected coulometrically. Carbon dioxide, e.g., Mocon Permatran C detects carbon dioxide in another dry gas by infrared (lightwave). Moisture vapor, e.g., Mocon Permatran W or dynamic water vapor tester measures moisture by a photoelectric sensor. alogens(e.g.use of carbons, HFCs), e.g., Krypton 85-high sensitivity is reported. icrobial integrity. Various procedures have been developed (under normal pressure and vacuum) to check whether highly contaminated liquid, gel, or media based material will grow-back or penetrate closure systems. Although these tests may be useful,variable results suggest that alternative methods to detect leakage are preferable. 6. Crack, pinhole, capillary detection. Conventional dyelvacuum immersion tests were widely used to check the seal efficiency of ampoules. These have largely been replaced by electrical conductivity and capacitance type tests. Typical equipment includes the Nikka Densok ampoule inspection machine, which employs a high frequency at high voltage. This distinguishes between good glass (which is a nonconductor) and areas of cracks or pinholes wherecurrent will flow between the inner and outer glass surfaces. Other machines use the principle of capacitance and dielectric constants where a material with defects(penetrating cracks) will show a higher dielectric constant than a solid non-defective material. 7. Acoustical. Tests can be based on sonic or ultrasonic energyof a gas escaping from a defect. Mainly used for checking pipes, pressure lines, and ducts, but under further investigation. 8. Thermal conductivity. These use a thermister bridge balanced against air and is subsequently upset if another gas leaks into the air. 9. Chemical tracer tests. Usedwith materials that can be detected by interaction, i.e., strong ammonia on one side, concentrated hydrochloric acid gas on the other (formation of white cloudof ammonium chloride indicates transfer and leakage). This type of test has been used for pinhole detection. 10. Thermocouple gauges. Mainly usedto detect a drop in temperature when solvent type systems escapeunder vacuum. Could also be usedto detect the presence of a warmer gas. ry ice (carbon dioxide) tests. Detection of loss. 12. Physical or mechanical assessmentprocedures. These have beenleft at the end of the list, not because theyare the least important, but because they involve the widest variety of factors, e.g., screwcaps are controlled by application torque or by removal torque. However, these changeaccording to material, design, environmental factors, etc., some of which may not have been fully investigated. (The assumption is often made that they have.) This equallyapplies to anyclosuresystem that relies on compression, interlocking, interference forces, etc. in order to make and maintain a “seal.” Thus certain forces can readily be measured, e.g., torque, force to push in a plug or pull out a plug type system, forceto apply a press over closure, forceto remove a press over closure. The variables associated with most of these systemstend to be quite large and complex.

ta~ility,and Shelf-Life

Whether the measured forces,torque, compression, etc., equate with an effectivesealwillvary according to- the quality of the materials employed and the perfection or imperfection of the surfaces involved between the two interfacing materials. In a paper on screw caps the author identified over 150 variables that could influence closure efficiency, 13. Visual inspection systems. The ability to inspect for and inspect out visual defects is constantly improving. These defects cover such factors as stones, glass inclusions, cracks, scratches, glass filaments,etc. in glass containers, and splits, burn marks, and molding deficiencies in plastics. These visual inspection procedures are also capable of high speeds, i.e., 400 units per minute and over. Automated visual inspection methods are gradually replacing the previous human inspection in which each unit was individually viewed by humans (frequently under 2x magnification) against a white or black background. Although not necessarily bearing a direct relationship with assessmentof closure integrity, these methods can eliminate likely suspect (imperfect) packs.

There are two basic test procedures (SCT and SFT) 1. Thesealcompression tester (SCT). The instrument measures the compression of the rubber-metal over cover combination, as a result of the sealing operation. This can be measured as a direct figure or as a percentage of rubber element compression (%REC). The latter is achieved by dividing the compression figure by the pre-seal thickness multiplied by 100. 2. The seal force tester (SFT). This instrument measures the static force exerted by the rubber element inthe sealed closure. The very first downward movement of the metal cap is quantified as the value of F at that instant and equates with the residual static force in the rubber component. This test is generally preferred to the SCT test. However, it should be remembered that a pack has other roles beyond that of basic protection, as shown in the definition givenin the introduction to this chapter. These include aspects related to identification, information, convenience, compliance, meetingbroad environmental needs, and fulfilling various performance functions. The latter include storage, distribution and production line efficiency, all of which have to be achieved in an economical manner. Changing any of these “performance” functions may bring about a need for a reevaluation of the shelf life achieved. ~istin~uishing between a minor change and a major change often creates problems. Closures, however, are critical to the total performance of the pack; hence they not only need special attention but are still one of the major reasons for a product recall.

There is an increasing demand for more guidelines to assist the selection and clearance of a suitable pack for any typeof product. Such information has to be submitted under the chemistry and pharmacy submission to a licensing authority.

Recent European (EU) Directives (2) have introduced a document covering the terminology to be used in drug submissions. (Notes for guidance EU/IH/ 3593/91 , e.g., the name for a rubber injection closure is a stopper, not a bung.) Allowance is also made for the use of Drug master Files (3)-see (EU/III/ 3836 /89. These may cover both product actives and excipients, and possibly packaging materials and components. EU Guideline I11I9090 (4) puts emphasis on packaging materials with particular reference to the useof plastics. It lists the information required in a drug submission under what could be considered as vague and sometimes difficult-to-define terms. Similar requirements are listed in a U.S.A. document entitled “Guidelines for Submitting ocumentation for Human Biologics” issued by the Center for Drugs and Biologics of the FDA in February 1987. A new document dated 1997 iscurrently under the review stage covered under CFR (Code of Federal Regulation) references. For example, a plastic parenteral container would typically require the following information: Name of manufacturer Type of plastic Composition, method of manufacture of the resin and finished container, plus a full description of analytical controls Physical description (size, shape) Light transmission (USP) USP tests Biological Physicochemical Permeation Vapor transmission test, if appropriate Toxicity studies, in addition to those included in the USP Compatibility including leaching, migration, plus sampling plans, acceptance specifications, etc. Although the above terminology is generally recognizedin the trade, much is not clearly or absolutely defined and hence is subject to “judgements.” The European Pharmacopoeia provides help in that it lists “permitted” plastics and also lists certain “permitted” additives. However, it includes in the list polymethyl methacrylate (PMMA) or Perspex (trade name)which the author has neyer found in contact with a pharmaceutical product (long term storage). A “permitted” additive, in the author’s opinion, does not mean that it is suitable for contact with a certain product until this has been established by contact tests. Guidelines also advise that polymers used must have basic “food grade appr~val,’~ otherwise expensive additional toxicity studies would be required. EU Food approval is covered by various directives, i.e., 901 128, amendedby 92/39 and others. These set specific migration limits as extracted under various conditions using a range of simulates, suchas distilled water, 3% acetic acid, 15% ethanol, and olive or sunflower oil. Section B of the directive identifiesmonomers and starting materials currently allowed, but which may be deleted if positive safety data is not supplied within. an agreed period.

Primary or immediate packaging 1. The nature of the packaging material, indicating the qualitative composition escription of the closure (nature and method of sealing) escription of the method of opening, and, if necessary, safety devices 4. Information on the container (single or multidose) and dosing devices 5. A description of any tamper evident closure and child resistant closure tics

This data should justify the choice of pack and include 1. 2. 3. 4.

Tightness of closure Protection of contents against external factors Container I contents interaction Influence of the ~anufacturingprocess on the container (e.g., sterilization conditions)

Specifications and routine tests, i.e., covering l . Construction, listing components 2. Type of materials identifying nature of each 3. ~pecifications,which may vary in detail according to product nature and route of administration. Routine tests are likely to include identification, appearance, dimensions, performance, bioburden, etc.

Listed under general and technical information for plastics is general information such as 1. Name and grade as used by manufacturer 2. Name of plastic manufacturer (parentals, ophthalmics) 3. Chemicalname of material ualitative composition 5. Chemicalname(s) of monomer@)used

aterial should have Food grade approval, otherwise additional toxicological data will be required. Technical i n f o ~ a t i o nrequired will include 1. Characteristics, general description, solubility in various solvents 2. Identification usually by infrared (the material) of the main additives and any dyes

3. Tests such as general tests, mechanical tests, physical tests, and extractive procedures using suitable solvents 4. Name of manufacturer/converter cti

Again, certain phrases are subject to interpretation. These can raise some challenging questions, i.e., l. If plastics are identified in the EP, does it mean that they are suitable for all ts? No (not without compatibility tests). plastic is identified witha list of approved antioxidants, are they approved with your product? No (not without further tests with product). 3. Stability and harmonization (ICH). Do the selected conditions of temperature and humidity providean effective challengefor the product and the pack? Someupplies have an identified and proven shelf-life? If not, reexamination date on the label?Often another assumption. importance does your company place on retention samples? Under are they stored and are they stored in suitable packs? Often not

asked. 6. Are retention samples kept of all packaging materials used both for tests and ongoing production supplies? Especiallyimportant during investigational tests. 7 . Are contact materials, such. as piping, mixing vessels,intermediate storage containers, packaging equipment, etc., given the same approval checks as packs? Usually, no. ve adequate controls been used in all testing procedures? Essential to trols for investigational tests. S attention to high science and high technology meant that simpler e overlooked? Perhaps this happens too often. his also questions the possible needof suitable “controls” at any stage during product pack development, i.e.,

rent starting materials /origins.Differing

methods of

ging scales of manufact~re fications in processes rent locations of manufacturing information is taken and recorded long samples are held prior to analysis Changing climatic conditions during processing eware “production manufacture” carried out over weekends, possibly with makeshiftstaff(i.e., manufacture made under production conditions ideally simulating the final production process to be used when the product is “sold”).

Remember the philosophy of thinking simply. Note: No list is all embracing. GLASS as a C O N T ~ O L When ! plastics are used, this answers the question whether it would have been more stable in glass. Note: There is a case for using sequentially numbered samples during the production of stability batches, as any unexpected differences from normal can then possibly be equated with the sequence of production, (i.e., any changing factors). AGI

Product-pack evaluation can extend outside the pack that contains the product to be sold or in which the bulk drug substance is to be held and distributed. All contact materials must have some form of investigation from discovery of a new chemical entity through all stages to product launch followed by outgoing regular stability evaluation. Contact materials may involvereagents used in analytical tests, vessels, piping etc, used in processing and any intermediate or temporary storage containers (at any stage). This observation and way of thinking will need to be extended inthe future. It also means that all investigational or feasibility evaluations may(will)have to be improved. This work is unlikely to reduce but become more intensive-hence a need for better “pack” checks will be inevitable.The ICH guidelines often assume that this work iscurrently done and is adequate, but there are and will be occasions where this is not the case. There is undoubtedly a need for better packaging research. There will be an increasing need to have a “competent overviewer,’ who surveys the total scene and discusses possible deficiencies with all concerned.This will require both “internal and external’’ partnerships to reduce risks and currently improve the total level of information-as part of a total data philosophy. Packaging technologists will therefore need to Recognize where “assumptions” were made in the past Improve their knowledge and information gathering Improve their powers of observation Effectively record all activities in detail Draw relevant conclusions Packaging in therefore a critical part of product appraisal and shelf-life and must be treated as such. The fact that it has been seen as the “Cinderella” of the industry must disappear and be replaced as a sound technological and scientific subject with a higher recognition as to its fundamental importance to pharmaceuticals, irrespective as to whether ethicals, generics, OTC, or veterinary products are involved, This knowledge mayalready be lessthan adequate with certain apparently well established drug substances, but each must be considered against the risk to userlpatient and the company. A key question iswhetheryouwould be satisfied by the data presented. Remember there are many ways in which “improvements” (the philosophy of quality assurance) can be achieved, but note even the most apparently logical arguments can contain flaws.

A typical “ass~mption’~ is that if all components are within the dimensional tolerances on a drawing then it is simply“within specification.” However,in certain cases, especiallythose related with closuresand containers-where it is within specification (minimum, mean, and maximum) may influencethe results. The author has seen few reports where dimensional detail has been included. For example, a pack with a screw cap is removed from a cycling cabinet, allowed to equilibrate with the laboratory environment, and analyzed. The results do not fit the normal expected pattern. What information could have possibly helped to explain the result-think about it! Could it have been a lack of attention to packaging and environmental storage detail? Try creating a list of variables, as once this list exceeds 20 factors then you know that you are beginning to really think!

The majority of pha~aceuticalproducts show good stability over a wide range of climatic conditions. minority, however,display instability that maybe related to a single challenge a multiplicity of challenges suchas variations in temperature, exposure to oxygen, light, moisture, etc. These challenges may be combated by storage under defined conditions (e.g., belowx°C,in a refrigerator, etc.) or in packs that act as effective barriers or restraints, to the criti eterioration factor(s). These cause and route of this is normally determined for t ug Substance and the Prodlopment” stages, hence it Osage Form, early in the “ should be relatively easy to find suitable storage conditions and a suitable form of “protective packaging.’, However, the full role of the pack extends beyond that of basic protection, as there are factors associated with overall quality, production performances, including assembly with particular emphasis on an effective closure systems, economics,etc., right through to ultimate use and final disposal. A packed product must therefore beseen asan effective marriage between the product and the pack, since the pack makes a major contribution to overall performance and to the declared shelf-life. The importance of the pack must therefore not be aving placed due emphasis e role of the pack, it is now possible lines on the future of the pack to consider the mostlikely impact of the I and packaging in general, and vice versa Conditions are often equated with a ‘6worst7’situation or theory both in the chosen testing conditions and in the selection of the lower linein regression analysisas representative of the product shelf-life. Variations within and between packs can also contribute to this situation by lack of effective evaluation and analysis withina stability program. An example of this can arise where a pack on test is removed from a hot storage condition and is allowed to cool prior to chemical analysis being carried out. At the higher te~peratureof stor ge the closure fit could be loose or looser and then tighten up as it cools, hence iving a false impressionof effectiveness, if only checked immediately prior to analysis. To cover this type of situation requires cap forces and physical appearance to be checked as follows: *

a. On immediate removal from the storage condition b. After conditioning to the laboratory environment c. Plus a possible further examination after product removal

Test (a) and (b) have to be carried out on separate samples. This makes an assumption that such work has not been evaluated during development investigational stages. With such a background any resultingvariations in analysis, then have a better chance of being coupled to possible pack variations or defects. It should be noted that there can be an assumption that there is significant data on plastic screw caps on plastic screw threaded bottles. This can often be untrue as the plastic used on both cap and bottle may vary, in addition to any dimensional detail. Experimentation is also complex as it needs to cover a range of torques (maximum, mean, minimum), immediate torques after removal from the storage condition, and torque after cooling or warming up (packs can also be stored at colder temperatures). This work needs to be supported by full dimensional details as the tolerances allowed within the specification may also contribute. Since torque evaluation is a destructive type of test, i.e., a cap can only be removedonce, a full statistically planned experiment can involve thousands of samples. To find out what happens if the cap isreapplied and further stored and removal is yet another experiment. The storage condition chosen under the IC guidelines, 40°C 75% R a very severe environment for a pack (try sittin it for one hour and se like it!) An indication of its severity has previously been given by reference to a typical crisp biscuit like Rich Tea, where shelf-like is related to its d of moisture, when a typical storage condition of 15 to 20°C with an is compared with storage at 40°C 755% RH can be a near equilibrium condition, should a product have an aqueous base containing dissolved “salts” in a permeable pack, hence little moisture loss might occur. Reducing the storage condition to t 40”C, which increases the humidity/vapor pressure gradient atmosphere and the aqueous product, will significantly increase the moisture loss, and the concentration of the active can increase. Again such a condition applies a severe challengeto boththe product and the pack, thereby giving an acceleration factor of a high order. In each of these examplesone must consider whether the product itself contains any freewater, which can exert a vapor pressure and introduce a contra-effect as temperatures increase. The potential for moisture loss or gain primarily depends on the nature of the product, the gradient between the inside to outside of the pack, the efficiency of the closure, and the permeability eral, materials that have a higher absorption to water have greater wever, differences betweenreal world climatic conditions and the ICH Guidelines must be constantly kept in mind. The guidelines normally only involve tests that employ the primary or immediate pack. The influence of the secondary or surrounding pack as required for display, warehousing, and distribution is normally not included in any tests. In the majority of circumstances these will serve to increase the shelf-life, particularly where effects due to moisture or gas permeation are involved. The surrounding packaging materials will also provide some “insulation” from the diurnal extremes in that the product is unlikelyto reach the lowest and highest temperatures or be fully exposed to the humidity extremes. This combination of insulation and enclosure, whereby certain possible exchanges/interactions are slowed down, can improve the product shelf-life significantly. This however,can conflict with the fact that testing under the current ICH Conditions may suggest that certain packs need better barrier properties than were necessary previously. For instance, the author has recently been approached

by three different companies whose blisterpacks have failedto provide an’adequate shelf-life either under the standard 25°C 60% RH and/or30°C 60% RH tests. This information is in conflict with some earlier products that are already being successfully marketed inblister packs (especiallyapplieswhere hard gelatin capsules are the product form). The ICH Guidelines also use one temperature and humidity at a time with no normal types of fluctuation other than those within the defined specification, i.e. f 2°C and f5% RH. As mentioned previously, most storage conditions experience the normal diurnal changes that typically arise during the day-night period. How the pack and the associated basicmaterials react to both static and fluctuating conditions requires more intensive study, particularly in terms of possible acceleration effects. Certain critical differencesbetween products and packs need highlighting, i.e., packs can dimensionally change according to the storage conditions, and certain properties can significantly changeas moisture content reduces or increases. These are usually independent of Arrhenius type orders of reaction. Materials can also change or “age” with the passage of time. The monitoring of possible surface changes can be especially critical and require the use of selected analytical techniques to examine the interface between product and pack. This trend towards more closely evaluating the product-pack relationship will inevitably vary according to the risk perceived to both the patient and the company and the use of “intensity of testing tables” willbecome more accepted and more critical. ~ackagingtechnologywill therefore need more science and a more scientific approach. This willneed to be supported by better validation and improved traceability. The fact that the committees creating the guidelines had little knowledge of packaging, as no experts weredirectlyemployed,will therefore generate a new era in packaging where attention to detail must inevitably increase and aspects not evaluated previously will haveto be considered. A s previously mentioned above, this will extend to more surface analysis and also involve estimations that quantify the purity and impurities in packaging materials more effectively. In conclusion, the advent of the ICH has undoubtedly more clearly defined formal stability testing, which with the use of matrixing and bracketing can lower potential costs. Although these costs have alwaysbeenhigh, and any savings can be significant, the guidelines are likely to put more emphasis on the level of testing required prior to formal stability, both in research and in development. Since the accumulation of data from the discovery of a new chemical entity to the ultimate withdrawal of a product (following a successful period of “life”) provides support both to product registration and its ongoing retention on the market, this work is also going to extend. The balance between saving on stability and spending more on general testing needs to be qualified and quantified as cost savings, long term, may be unlikely. Improving the data available at the earliest stage will, however, be commensurate with the earliest launch date. This will apply both to research and development and cover suchfactors as faster development and better analytical methods, improved chemical substance screening, clearer understanding of the product form required, including release characteristics, extension of the use of surface analysis techniques, etc. Areas which require special attention (especially wherenot covered by the ICH ~uidelines)include

A. Influences due to storage (stacking, drops, compression) and distribution (impacts, compression, vibration, puncture, etc.). Note top compression, and the influence of vibration, can be particularly critical. Cycling conditions as found over various parts of the world (i.e., those in y conditions are significantly different from hot/dry and hot/ humid areas of the world). Just testing under the static ICH Conditions is unlikelyto be adequate. C. Patient use, testing and finaldisposal. Again, thismayneed to cover temperature, humidity extremes, and cycling conditions as indicated in Ref. 2. However, possible patient misuse or abuse maybelesseasy to define, i.e., will patients blow hot moist breath through certain types of ‘Yung” inhalation devices prior to use in order to “clean7’off any possibleadhering debris? Although checking for suchpossibleactivitiesmay be uncovered through patient questionnaires, patients can be oblivious of what they do, hence patient observation may be an essential part of any evaluation. Advice on disposal may also be required, especially whereunique materials are used, or large quantities could create a disposal problem (e.g., large quantities of neutral Type I glass can cause severe problems if incorporated into other types of glassware, i.e., types 11, 111, and W). Potent low-dosage drugs can also become us substances when shipped in bulk. Improve evaluation and quantification of pack details and any associated iations. This basically means that knowledge of raw materials, pack conversion processes and components, and assembled packs must be more effectively assessed and then held in databases. Unfortunately many experiences of the past tend to be anecdotal and are lost as people retire. Achievingeffective control of such basic detail is particularly important for Third World countries, who frequently suffer from the sophisticated world’s often long lost or forgotten earlier problems. Such information also tends to be edited out of books, as they are seen as outdated unnecessary information. Creating special books for the Third World also tends to be economically nonviable, as there is often no guarantee of sufficient sales. However, it can be said that attention to packaging detail, including levelsof “purity” and traceability, remains a weak link in the general clearance of a product-pack. Checking of the pack, whensamples are withdrawn for stability assessment, is one further aspect, which requires special consideration. It should be noted that packs do not necessarily follow any normal rules of degradation, e.g., polymersexpand and contract according to conditions of temperature and, to a lesser extent, humidity. E. How batches of product are pared for formal stability also needs special attention with full coverage ofa1 ails. However, it should be noted that materials used in any type of test from initial research to the end of development should undergo a full QC clearance assessment (calledQC plus by the author). Until a “final” specification (there is no such thing-all specifications should be subject to constant review) is made, there are many factors that may need quantifying and qualifying prior to them being deemed either a necessary or an unnecessary part of the “ultimate” specification. Finally, the challenging conditions advised under the ICH Guidelines (no real conditions are at a constant 40°C 75%/ 15% RH or even 30°C 60% RH) may infer that more protective packs maybe required inthe future. Whilst the author suggests

that general feedback from the market place does not support this observation, the means by which barrier properties could be improved needs, at the very least, to be identified and considered. The basic options may therefore involve choosingfrom the following: F. A better barrier material, e.g., glass or metal to replace plastic, a plastic bottle to replace a blister pack, etc. C. Use of the same or similar materials, with an additional barrier coating or a more effective barrier coating. acing one or more ad~itional“overwraps” around the primary pack or e barrier properties of the secondary materials. ognizing and proving that the secondary packaging materials can add to the shelf-life of the product. J. Accept a reduction in shelflife.

It would appear reasonable that the risk (to patient/user and company) could and should bear some relationship to the intensity of the tests involved in the clearance of the materials from which the primary or immediate pack is made. Intravenous (IV) solutions head this list as with the large volumes involvedas even a small level of “extractive” will be rapidly in contact with the most sensitive body organs, i.e., heart and brain. e.g., lOOOmL extractive 0.1% = 10 gm (IV solution) 20 mL extractive 0.1% =20 mg (nasal spray) Although this example has been oversimplified, it should confirm a need for different intensities of testing according to the product category/route of administration. This will also be influenced by A. The product-to-pack contact area . The nature of the product (che~ical-physical characteristics) C. The contact time and temperatures (etc.) of storage and the nature of the extracted constituent The list below suggestsa product “intensity of testing” order, which is headed by sterile products. However, it must be stressed that the position in this order may change according to any special product characteristics, contact area, contact time, te~perature,etc. The list is therefore given to create a greater awareness of how “risks” to “benefits” might be considered in packa~e-productclearance. Some additional “other factors” are i uded as a separate list.

arge-volume parenterals (LVPs) IV solutions rigation solutions ialysis solutions a. Small-volume parenterals (SV

IV additives Intravenous Intramuscular Subcutaneous Intrathecal Etc. b. Implants phthalmic products 1 and unit dose d. Nebulization solutions (may be sterile or preserved) roducts (may be sterile or of low bioburden) dministered to the lungs Transdermal patches Vaginal and rectal products uccal products Local and topical products Liquid oral products Solid oral products Note: Lists are not 3. Intensity of Testing a. Patient type,i.e b. Treatment point or period c. Volume of product to pack contact area d. Nature of products, tend to be more extractive than. water based materials (rub ics) e. Storage cond ditions and “others” f. Storage perio g. Processing factors., e.g., terminal vs. aseptic processing h. Type of pack, i.e.., bulk pack for dry substance ulk pack for sage form for dosage form (unit of use pack) onsider a. Type of product-ethical (NCE) Ethical (product extension) Ethical reformulated modified process Change in pack-major or minor TC eterinary b. Packsizevariants-volume,weight, or number c. Storage period and storage status: unopened or opened-including ns”contro1led (simulated) (“others” and “real?’); ling (simulated); ambient (actual) ge and transit conditions ods of use-Instructions-use-misusey abuse

1. Beware of assumptions 2. The need to quality control all materials used in tests/evaluations by “QC plus” 3. The need to quantify all activities-keep good records 4. The importance of traceability-background information

The issuance of EU,WHO, and ICH Guidelines for formal stability programs makes many assumptions. For example, it is assumed that a company builds up in-depth knowledge on any new chemical entity or drug substance before it is put on a formal stability program, so that the chance that it will fail or not live up to expectations, when stability tested, should be remote. In the case of an established drug (as used for OTC products or generics), initial confidence is achieved by literature searches or cross-references to existing data. This assumesthat the drug substance is synthesizedby the same route, but if this route changes, then additional extensive testing may be required. Formulated products follow a familiar pattern. With ethical products the initial stability of product and pack(s) are achieved by accelerated and then longer term tests on the product, as associated with safety testing on animals, human volunteer studies, initial and ongoing clinicalevaluations and general investigational studies, which are likely to include reformulation challenges, pack-product investigational studies, etc. Through allthese studies the product-pack has to be supported by a shelf-life calculation, based on defined specifications. The various guidelines therefore make assumptions that this type of work, often covered under the heading of a “total data philosophy,” has built up highconfidence at all stages of research and developmentcoveringall the materials employed. Part of this confidence r e q ~ i that ~ ~ sall incoming materials related to product and pack undergo a full QC examination-what the author calls “QC plus,” since it is on this information that a major part of subsequent specifications are developed. Knowing exactlywhat is being usedin tests and evaluations generally, requires greater in-depth knowledge than what will be required at the final marketing stage, i.e., at early stages onetends to quantify all possible factors, as those that are most critical and need includingin the final specifications will only emerge following experimentation and experience. This is particularly important where the product and/or pack are totally new with little or no prior historical experiences. This needto build up information must be applied at all stages, being based on effective

1. Observation/information 2. Recording in detail 3. Drawing anyrelevantconclusions This should include how products are made, how packs are assembled, with what materials, how evaluated, etc., thereby being covered by the broad term “traceability,” i.e., the ability to trace the history or background of any material or operation. The challenges that achieve this information invariably involve the use of such interrogative pronouns as how, where, what, by whom, why, which, when, etc., as part of the recorded information required. One also has to decide from this detail which answersare most relevantto a particular situation. The guidelines therefore make assumptions that such work has been effectively completed

~tability,

al~ation,

and

before formal stability testing is started on the drug substance and the product. The more critical aspects of this work should bewritten up as formal documents, which may form both part of a regulatory submission and part of the justification of selecting a product in a specific pack form. In the investigative work that supports the above, it may be necessary (and indeed advisable) to use both ICH and other challenging conditions, whichshouldinclude adequate controls. These controls may involve selected conditions (many products, though not all, are more stable in a refrigerator) and other types of packs. If a product exhibits somedeterioration in plastic, a glass control answers the question “Would it have been more stable in glass?”, hence a glass control is advised in early investigational stages, etc., e.g., a plastic screw-capped bottle containing a product shows discoloration of bottle threads when the cap is removed: it raises the question whether this is product or pack material related. Controls could be samples of pack with no product, plus separate retention samples of bottle and cap, as this enables the source of the problem to be quantified. Packs may also have to be stored in various positions, i.e., upright, on the side, upsidedown, as these may also influence product stability, etc. Product evaluation and stability can also relate to many other possible variables, i.e. People and their responsibilities/functions Materials employed Equipment-instrumentation-machinery Documentation Environmental conditions and their control, as used for tests. Environmental conditions of any operation or process f. Services, and their control g* Facilities, and their general quality h. General legislation, including packaging and wastedirective,child resistance, etc.

a. b. C. d. e.

Each of these factors could be in-depth discussed as part of the aspects of R and any area we can be guilty of making assumptions, as toooften all of us believe that we always do an excellent job and that there are no problems associated with our equipment, instrumentation, services, facilities,etc. Even with today’s emphasis on validation, deficiencies can still arise. Moisture loss or gain in a climatic cabinet can vary according to the type of shelving used (solid or perforated) or whether samples are tightly packed or loosely packed (have air spaces around them), and this is one area which is rarely investigated. Such conditions are also difficult to validate fully, as conditions may change dependingon whether the area is full, part full, or nearly empty, the nature of the material therein (absorptive or nonabsorptive of moisture, heat, etc.) and other factors. Such simplefactors can be overlooked as the sophisticated analytical methodology now available usually has greater interest. The useof regression analysis, where the worst line of fit may be used for the shelf-life, can often be influenced byoutliers, which could be packagingrelated, e.g., tightly packed samples will show slower moisture loss or gain when sampledat 3 or 6 monthly intervals at the beginning of a stability test, whereas the same when sampled annually (year two onwards) from a less tightly packed stock may showa different pattern of moisture exchanges.

~xperiencehas indicated that most people would like to ask the question: “Is there a setofrules or guidelines to which I can work with reference to product-pack stability?,’ Although the answer to this is now a broad “yes,” there will be certain exceptions, i.e.,factors, that are special to each product-pack-administration system that needs identification. This means that specifications and the relevant test procedures cannot be totally set by a fixed approach. What is relevant will therefore vary from product to product, This thinking also applies to what is generally termed “product quality,’?including “elegance” as this also will vary according to the product, its manufacture, and the situations in which it is stored, transported, displayed, and used. Guidelines are also under constant review where areas have to be updated as new ideas of information, recognition of points of weakness, etc., become identified. In these circumstances,a product registered in territory A five years ago may havean area of deficiencythat has to be investigated and cleared beforeit can be registered in another territory. The standards of yesterday are therefore unlikely to satisfy totally those of tomorrow.

There are many process procedures that can modify the properties of plastic. The more obvious ways in which plastic performance can be in~uencedinclude 1. Changes in the solubility of the absorbing materials in the polymer 2. Changes in the solubility of the absorbing materials in the surface layer 3. Changes in the surface conditions 4.. ~ o b i l i t yof permeate within the plastic 5. Opening up a molecular structure by swelling (increases permeation) vement of internal constituents to the surface (or vice versa) 6. anges in mass or density 7. olecular changes due to degradation, including chain fission 8. 9. Changes in particle size or dispersion of constituents, etc.

If any of the above or combinations of the above occur, changes mayarise to physical,physicochemical, or chemical properties, e.g.,in terms of diffusion, permeation, or migration/leaching. Thorough retesting is, therefore, sometimes essentialto detect changes between the and ‘‘unprocessed” (as supplied initially) plastic. It has also to be recognized that stored plastics may change or “age” with the passage of time prior to their use to pack a product. This can mean that greater attention needs to be paid to how materials are packed, stored, etc.

United 1. States Pharmacopeia (USP231NF18). Rockville, MD: United States Phar~acopeialConvention, 1959-1963. 2. European Union Guideline, Notes for Guidance EIJ/III/359/91.

3. European Union Guideline, Drug Master Files EU/III/3836/89. 4. European Union Guideline, Plastics111/9090/90. 5. United States Pharmacopeia (USP23/N~l8). Rockville, MD: United Pharmacopeia1 Convention, 1959-1963, pp. 1781-1790.

States

This Page Intentionally Left Blank

~ h a r m a c e ~ t i c a~l o ~ ~ u l tKalamazoo, a~t, ~ichi~an

1. Introduction

516

2. Stability Testing and Product Development 2.1. Stability testing of formulations for preclinical safety testing or toxicology testing (preclinical stage) 2.2. Preformulation stability testing ofnew chemical entities (pre-IND stage) 2.3. Accelerated and normal storage testing of clinical formulations (IND stage) 2.4. Stability testing of early formulations, formulation mixtures, and packaging materials evaluation (product development stage) 2.5. Long-term primary stability testing of drug substance and final marketable formulations in proposed packages for marketing for product registration and approval (NDA stage) 2.6. Stability testing of production batches post regulatory approvals (approved product stage) 2.7. Stability testing of revised products (revised product stage)

517

3. Development of the Stability Testing Function 3 1. Written stability testing program 3.2. Stability protocol and commitment 3.3. Standard operating procedures 3.4. Training of personnel 3.5.Selectionof batches and samples 3.6. Stability storage test conditions and stability roomslcabinetsl chambers 3.7. Thermal cycling 3.8. Photostability testing

518 519 520 521 521 522 523 523 524 524 528 530 53 1 532 534 534

3.9. 3.10. 3.1 1. 3.12. 3.13. 3.14. 3.1 5. 3.16.

Container/closure systems Testing frequency Stability indicating analytical methodology and validation Stability of reconstituted products Dosage form considerations Bracketing Matrixing Statistical analysis and evaluation of data nt of Stability Information Using Computers

536 536 537 537 537 539 540 541 543 548 549 549 551

Stability testing is an integral part of the pharmaceutical development process.It is routinely performed on drug substances and drug products. Drug substances can be derived from chemical synthesis,from plant or animal sources, or from biological, biotechnology, or recombinant DNA technology. Stability testing is carried out on drug products from chemical or biologicalsources, prescription drug products for human health and for veterinary use, over-the-counter drug products, and medical devices. Stability testing principles are applicable to all of the above types of drug substances and drug products. Specific unique requirements covering each of these areas exist. A thorough description of the stability system covering these areas is beyond the scope of this chapter. This contribution primarily deals withthe stability of pharmaceutical new chemical entities and drug products for human health. The chapter relies on principles and concepts discussedin the recent draft of stability guidelines (FDA, 1998b) published by the Food and Drug Administration in the United States. It is recognized that the FDA guidelines are in draft format, which have not been finalized as of this writing. It is not known whether the guidelines will have been finalizedby the time this monograph is published. It is this author’s assumption that there will be changes in the final guideline from the current draft, but it is my expectation that the current draft guideline will retain most of the key contents without significant modifications. In addition, the guidelines covered e Tripartite Report on the International Conference on armoni~ationare cable to the chapter contents (IC The primary purpose of stability testing in an industrial setting is to provide supporting evidence on stability behavior of chemical or biological entities as well as tostudy the sta~ilitybehavior of pharmaceutical drug products. Stability testing as a function of time against a variety of environmental factors such as temperature, idity, light, and combinations of these parameters is critical for the establishment of recommended storage conditions, retest periods, expiry dates, and shelf lives of pharmaceutical products. Stability behavior in broad terms refers to overall quality of drug ingredient or drug product in terms of strength, purity, identity, safety,

apparent degradation, physical or biological changes, and their effect on biological performance of drug products. any factors drive the process of industrial stability testing for pharmaceutical products. Theseinclude the generation of background information for the development ofnew products, the support of manufacture and distribution of pharmaceutical products in different regions of the world, the development of suitable packages to insure the quality, strength, purity, and integrity of the product, the satisfaction of the submission of stability information requirements for regulatory agencies around the globe, the determination of shelf life,and several other factors. chapter is divided into three sections. The first section, “Stability Testing t Development,” covers general aspects of stability principles involved in the product development process. It provides the objectives of stability studies during various stages of drug substances and drug products as they evolve from the discovery phase to clinical formulations, to drug product for marketing, and revised product post approval. The next section, “Development of the Stability Testing Function,,’ covers the development and establishment of the industrial stability testing function, It includes various activities and associated regulatory guidance information related to written testing programs, standard operating procedures, stability protocol generation, testingfrequency,selection of batches, matrixing, bracketing, and other factors that must be addressed in establishing the industrial stability testing function. A discussion ofthe documentation required to minimize the chances of regulatory citations is included. This section provides information related to stability test conditions and stability rooms, cabinets, and chambers as well as a table with a list of suppliers and manufacturers of constant temperature/humidity rooms, cabinets, and chambers in the United States. In addition, principles for photostability testing and a list of the suppliers and/or manufacturers of the photostability testing chambers are included. The third section, “~anagementof Stability Information Using Computers” provides an account of the principles involved in the development of computerized stability information systems. Factors involved in the design of stability systems from data operations, performance factors, systemsanalysis,systemsdesign, validation, maintenance, and other principles are included. In addition, a table with a list of developers and suppliers of stability system software and laboratory information ~anagement systems for pharmac~uticalapplications is included. *

Stability testing function is an evolutionary concept covering the life cycle ofpharmaceutical product development. The objective of stability testing varies during various stages of product development. For example, during the early discovery phase, the primary focus is to generate stability characteristics of a chemical ingredient or biological entity, which will be helpfulin the design and development of drug or biological products. uring later stages of product development, the goal of stability testing is to establish shelf life for formulations packaged in a final package intended for commercial introduction. Stability testing principles can be subdivided into various

lvani

stages of development. The objective of stability testing differs in each stage of development. The following sequence describes several stages of drug development during which stability characterization should be carried out. Stability testing of new chemical or new molecular entities (discovery phase) Stability testing of formulations for preclinical safety testing or toxicology testing (preclinical stage) Preformulation stability testing of new chemical entities (pre-IND stage) Accelerated and normal storage testing of clinical formulations (IND stage) Stability testing of early formulations, formulation mixtures, and packaging materials evaluation (product development stage) Long-term primary stability testing of final marketable formulations in proposed packagesfor marketing for product registration and approval (NDA stage) Stability testing of production batches post regulatory approvals (approved product stage) Stability testing of revised products (revised product stage) Stability testing of new chemical or new molecular entities (discovery phase) During the discovery phase, stability testing of a new chemical entity or a new molecular entity is requiredto help selectthe most satisfactory chemical or molecular entity possessing the right pharmacological, toxicological, and pharmaceutical profiles. The pharmaceutical profile is mostly focused towards the optimum chemical and physical stability characteristics, a good preformulation profile, and satisfactory manufacturing potential. The emphasis initially is not only towards selecting the right chemical entity but also towards selecting the appropriate physical form (for example a satisfactory polymorphic form if applicable, or a physical form with the most desirable handling behavior), and, where applicable, the base, salt, ester, hydrates, solvates or other forms with optimum stability. It is not uncommon to find vast differences in stability characteristics of different salts of the same chemical entity. In addition, preliminary information regarding particle size, distribution, crystal shape, crystal habits, ~echanicalproperties, and specific surface area can play key roles in affectingthe stability, manufacturability, and biological performance of the chemical entity. These studies help establish the boundaries within which one must operate to design formulations for toxicology and for initial clinical testing. En the current competitiveaccelerated drug development environment, early physicochemical investigation has to be completed not only in the shortest time possible but also with limited quantities of drug substance.

For a new drug to be evaluated for preclinical safety assessment, often in several animal species, it must be formulated in a dosage form or a delivery system that will deliver the drug in a way that maximizes the availability of drug at the site of action. The development of early dosage forms for preclinical testing and eventually for early testing in humans requires an extensive stability evaluation and interplay of physicochemical, biological,and dosage form considerations. Good laboratory practices (GLP’s) dictate that all substances or formulations administered

to animals during preclinicalsafety testing remain stable for the toxicological t eliminary stability testing on all formulations ried out using indicating assays in accordance with toxicology for~ulationsare not adversely affected d ~ r i n gtox many toxicologyformulations are suspensions, homogeneit and conte are critical stability parameters, which must be evaluated. good homogeneity to deliver ~niformdosing and sboul stability, especially ~ i n i m a settling l ten~encywith good resu m stability testing for alltoxicology formulations usedin requires that an entrance assay prior to the initiation of toxicol and an exit assay, must be performed at the end of the studies.

including the identi~cationand quan

during normal storage or under accelerated temperature conditions of long-term stability testing, they do help establish the worst case scenario., and importantly they help in the development and validation of stability indicatin cedures.

egulatory guidelines in the United States as well as in other parts of the world require that formal stability testing of early formulations for human use must be initiated. For example, the regulatory requirements for ~nvestigational New drugs ’S) state that information sufficient to support stability of the drug substance during the toxicological studies and the planned clinical trials should be available. The goal of these studies is to elucidate identity, quality, strength, and purity, and to help establish the retest period for drug substance. ~imilarlyfor clinical formulations, the agency requirements state that the informatio~ to assure the product’s stability during the planned clinical studiesshould be available. The stability data generated during this early phase isused to establis~ storage conditions and specifications. Stability information for submission during this phase should include a brief description of the stability study, test methods for monitoring the stability of the clinical formulation in the proposed container/closure, and tabular stability data of the representative batches. Accelerated and normal storage temperature testing of drug substance and for clinical formulation must be initiated prior to the initiation of clinical studies. The goal of these studies should be to generate information to insure that the clinical formulations are likely to remai table during the planned clinical studies. Genat least one-month stability data at accelerated erally at the time of filling an I conditions for a single batch clinical formulation should be available‘ This is consistent with the Food ug Administration’s policy,which states that the regulation does not prec ponsor from conducting stability tests on an investigational drug product concurrently with clinicalinvestigations of the product. ever, the agencydoesexpect th the timeclinicalstudieshavebegun, the sor will have submitted to the at least preliminaryevidence (obtained from accelerated studies) to show that the product is likely to remain stable for the duration of the study. Further insight regarding thegradual natureof the investigationcan be gained from the FDA’s comments regarding INDs, which state “the regulations at 21 CFR 3l2.~3(a)(~)(i) emphasize the graded nature of manufacturing and control information. Although in each phaseof the investigation sufficientinformation should be submitted to assure the proper identification, quality, purity, and strength of the investigational drug, the amount of informationneeded to make that assurance will vary with the phase of the investigation, the proposed duration of the inv~stigation,the dosage form, and the amount of information otherwise available. For example, although stability data are required in all phases of the IND to demonstrate that the new drug substance and drug product are within acceptable chemical and physical limits for the planned duration of the proposed clinical investigation. If very short tests are proposed., the supporting stability data can be correspondingly very limited (FDA, 1995a).

The goal of the pharmaceutical scientist isto design and develop a dosage form that is stable (for the entire shelf life or until the expiry date), that contains the precise amount of drug that will be delivered in the most available form, and that can be manufactured on a large scalein an economic manner to meet regulatory approvals and market needs. Stability testing must be performed to ensure that the drug products not onlyremainchemically stable (maintaining their fully potency) but also retain satisfactory physical stability attributes in terms of elegance, appearance, lack of discoloration, as well as in-vitro and in-vivo performance characteristics (for example, dissolution). uring product development stage, stability testing of several early clinical formulations leading upto marketable formulations should be performed. Also a systematic evaluation of container/closure systems and packaging systems is initiated to select an acceptable kaging materials evaluation makes use of the preformulation infordrug substance as well as stability evaluation of clinical formulations. The objective during this phase is to identify the stability limiting factor(s) and he1 establish shelf life for products. Since manyproducts go through several distinct steps during manufacturing, it is very commonto have intermediate granulation, powder mixes,solutions that need to be stored prior to the next manufacturing step. Stability evaluation of intermediate mixtures that are likely to remain in the intermediate state for more than a few hours should be performed. For example, whena mixture of active and inactive ingredients or a granulated mixture waiting to be compressed into tablets or to be filledinto capsules isproduced, a short term stability evaluation in an appropriate packaging system needsto be performed. Similarly, if the finished dosage forms are stored in interim containers for an extended period of time prior to fillingin marketed containers, real time stability data should be generated to justify the length and condition for acceptable storage. ften this type of short-term stability evaluken during pilot plant scale-up or validation operations. Gener of finished product exceeding six months is not appropriate. force for this consideration is that the expiry date is computed from the date of quality control release of the batch, but no later than 30 days beyondthe date of manufacture, and is independent of the packaging date.

arly stability studies, preformulation studies, and ongoing development studies with a drug substance help establish the formal stability program for generation of sta~ilitydata for registration applications. Critical stability characteristics and stability results should be usedto determine the appropriate storage conditions, internal use, and expiration dates for drug substances being sold e reader is referred to the detailed discussion in Section 3.5 on the selection and n u ~ b e of r batches. Stability of these batches should be d e t e ~ i n e d using stability indicating assay methodology. Testintervals should follow the mini-

mum frequency as covered in Section 3.10. A stability program should include a commitment to follow the stability of a number of batches annually. The stability of drug substances should be evaluated in containers that simulate the storage container used for warehousing or the container used for marketing. For example, if the drug substance is stored in polyliners within fiber drums, stability samples should be stored in smaller polyliner bags of the same material and smaller size fiber drums or containers of similar or identical composition to the large-scale drums. Care should be exercised in selection of the size, surface-to-volume ratio, and air space, as they could affect the relative stability, possibly because of differences in relative humidity or because of specificinteractions with polyliners. Changes in ~anufacturing site, materials, or manufacturing process require careful evaluation, as they could affect stability. ience gained from clinical formulations of drug product during earlier the knowledge of stability behavior of drug substance help establish program for the drug product for registration purposes. Critical stability characteristics and stability results should be used to determine the appropriate itions, test intervals, and batches for the product registration stability should be noted that strength in terms of chemical stability is not the only criterion of drug product stability. rug products must maintain various c h e ~ i c a land physical prop~rtiesto erve the effectiveness, safety, elegance, and overall acceptance of the dru operties such as physical appearance,, crystalline form, particle size, solubility, disintegration rate, dissolution rate, pH, lity, viscosity, palatability (taste and odor), homogeneity,resuspendibility, stability after reconstitution may be stability related and thus require testing of specific storage conditions and limits. n addition, tests may also etermine the absence or presence of harmful degradation products. ay dictate an upper limit on the level for degradation may be a stability limiting characteristic. Stability testing parameters should include those characteristics that are likely to influence quality, strength, safety, and/or efficacy. The reader should consult the detailed discussion concerningthe selection and the number of batches as included in Section 3.5.

he goal of the stability evaluation program during this phase is to confirm and/or extend the e~pirationdating of the drug product that has been approved. In addition to the continuation of the monitoring of primary stability batches, a commitment should be made that long-term stability for the first three post-production batches rated using an approved stability protocol to justify the confirmation n of the expiration dating period. The commitment also includes monitoring of the stability of annual batches after the first three production batches. The commitment actually constitutes an agreement withthe Food and istration indicating that the plan outlined in the product registration document will be strictly complied with. The commitment also requires submissionof stability data at periodic intervals andlor annually as specified in the application. commitment mandates that any batches that are found outside of the specifications approved with the product will bewithdrawn from the market. If the deviation from

the approved specifications is observedto be a single occurrenceand is not expected to affect the safety and efficacy of the product, regulatory guidance allows notification to and discussionwith the approving regulatory divisionwith detailed justification for distribution of the product.

and size of dosage forms, changes in manufacturing site, changes inmanufacturing process, and so on. They may be externally driven, for example, deleti regulatory restrictions on formulation in~redients,formulation changes. ciple in the evaluation of stability for revised products is partially to revalidate the test method and confirm the applicability of stability indicating assay methodology or undertake modifications of the assay method. st post-approval changes for a revised product require some type of stabation and comparative stability along with a commitment to monitor the stability of the first one or three production batches of drug substance or drug product and annual batches thereafter. The has published a detailed account the type of commitment through , 1995b). The type of stability data required for submission rangesfrom no additional stability data for a manufacturing site change within the same facility withthe same equipment to 3 months of comparative accelerated stability data and available long-term on three batches of drug product with the proposed change.Likewise the stability commitment ranges from none to the first three production batches and annual batches thereafter on long-term studies. Overall, the goal is to apply good scientific principles and generate comparative stability data at accelerated conditions and long-term stability data with the proposed changes and a commitment to follow up long-tem stability with the production batches.

ev~lopmentand establishment of a stability testing function in an industrial setting g stability studies requires careful consideration of several areas of interest. require alot of resources;they add alot of expenses to development cost and require a long timefor the generation of stability data. A s discussed in the previous section, many batches of drug substance and drug products are put on long-term and accelerated stability testing from the initiation of stability studies during the discovery phase to clinicalphases and finally the primary stability ultiple batches are evaluated to ensure that a product will consistently remain within specificationsfor its entire expiration dating period. Stability studies are routinely conducted to investigate the effect of strength, package, batch, storage condition, and stora e time on the stability of the dosage form. For primary stability data generation a typical study will have three batches, several strengths, several packages, and several storage conditions. Thus if every

batch by strength by package combination is tested for every storage condition, a substantial expense is involved. To reduce expenses, a bracketing design is commonly used;for example onlythe smallest and largest bottles are tested. In addition, the application of sound statistical principles by using fractional factorial-type matrixing designs for conducting the stability studies can substantially reduce the amount of testing required without compromising the outcome of results (Fairweather et al., 1995; Nordbrock, 1992). For complex stability studies involvin many variables, a combination of bracketing and matrixing design can be used to gain some further economies in testing. Stability programs need to be well managed for efficient and effective operations. It is critical that a thorough review of ongoing stability studies be undertaken on a periodic basis to reduce and/or control costs. Establishment of the stability testing function requires an extensive development of documentation to maximize the compliance and to minimize the chances of regulatory citations. It requires setting up ~ u l t i ~ stability le test conditions and stability rooms/cabinets/chambers.A table with a list of suppliers and manufacturers of constant temperature/ humidity rooms/cabinets/chambers in the United States is included(Table 1). In addition, a list of the suppliers and or manufacturers of the photostability testing chambers is included (Table 2). The following isnot intended to be a comprehensive list. It covers severalcritical items that must be addressed. Specific guidelines for many of these exist in the guidance documents issued by the Food and rug Administration and the

(Code of Federal egulations) state, “there shall be a written testing program designed to assess the stability characteristics of drug products. The results of such stability testing shall beused in determining appropriate storage conditions and expiration dates. The written program shall be followed and shall include .,.” is critical for the establishment of the stability testi must address key issues suchas the selection of ba indicating assay methodology, the testing freque of container/closures, and the stability of recon cedure should also include the review and appro for storage of raw stability data, and the peri addition, the storage of samples, the procedu results, review, action plan and sign-off proce storage malfunction, and many other issues S line shouldbe part of the written documentati tion in documentation. Every action perform function must be documentedin a SOP o agencies maintain that activities may not be claimed as having been p~rformed without documentation.

enerally a stability protocol should be generated for every batch that is being followed on long-term stability testing. The stability protocol should record the purpose for co~ductingthe stability test, the method used, the testing frequency,

anufacturers of Stability Testing Rooms, Chambers, and Cabinets, Environmental Chambers,Humidity Chambers Name

PhoneAddress

E-maillweb lfax

site

Absolute Control Systems

5168 Parfet Street, Unit G Wheat Ridge, CO 80033

Phone: 303-420-8600 Fax: 303-420-8692

E-mail: abso1ute~minet.com Web site: http: 1lwww.absolutecontrolsys. corn

ARS Enterprises

12900 Lakeland Road Santa Fe Springs, CA 90670

Phone: 562-946-3505 Fax: 562-946-4 120

E-mail: [emailprotected]

Atlas Electric 414 N Ravenswood Avenue Device Co. Chicago, IL, 60613

Phone: 773-327-4520 Fax: 773-327-5787

E-mail: mmacbaeth~atlas-

Barnsteadl Themolyne

2555 Kerper Blvd. Dubuque, IA 52001

Phone: 319-556-2241 Fax: 319-589-0516

Benchmark Products Inc.

531 Bank Lane Highwood, IL 60040

Phone: 847-433-3500 Fax: 847-433-3545

E-mail: ~n@benc~arkprods.com

Biocold 1724 Westpark Center Environmental Fenton, MO 63026 Inc

Phone: 314-349-0300 Fax: 314-349-0419

Web Site: http: l/members.aol.com/biocold

Brinkmann Instruments, Inc

One Cantiague Road P.O. Box1019 Westbury, NY l 1590-0207

Phone: 516-334-7500 E-mail: [emailprotected] Fax: 516-334-7506 Web Site: htto: l/www.brinkmann.com

Caron Products & Services, Inc.

PO Box 71 5, Products Lane Marietta, Ohio 45750

Phone: 740-373-6809 Fax: 740-374-3760

E-mail: c a r o n ~ f r o ~ e t . n e t Web Site: http: 1l~w.caronproducts.com

Clean Air Solutions

501 1 Falling Leaf Trail North Syracuse, NY 13212

Phone: 315-452-4609 Fax: 315-458-0162

E-mail: [emailprotected] Web Site: http: /Iwww.cleanair-solution.com

Cole-Parmer Instrument CO.

625 E Bunker Court Vernon Hills, IL 60061

Phone: 847-549-7600 Fax: 847-247-2929

E-mail: [emailprotected] Web Site: htto: l/www,coIeoarmer.com

612 W. Stutsrnan Street Controlled Environments, Pembina, ND 58271 Inc.

Phone: 800-363-6451 Fax: 204-786-7736

Constant Temperature Control Limited

220 Industrial Parkway South, Unit 22, Aurora, Ontario, U G 3V6, Canada

Phone: 905-841-7749 Fax: 905-841-1669

DJS Enterprises

110 West Beaver Creek, Unit 14 Richmond Hill, Ontario, Canada LIB 1J9

Phone: 905-764-7644 Fax: 905-764-7654

Edge Tech Moisture Humidity Instruments

455 Fortune Blvd. Milford, MA 01757

Phone: 800-276-9500 Fax: 508-634-3010

E-mail: [emailprotected] Web Site: http: /1www.edgetech.com

lvani E-maillweb lfax

PhoneAddressName Electro-Tech Systems Inc (ET9

3 101Mt Carmel Avenue Glenside, PA 19038

Phone: 215-887-2196 Fax: 215-887-0131

Environmental 510 E Washington Street Phone: 800-321-6854 Growth Chagrin Falls,OH ~022-4448Fax: 440-247-8710 Chambers Environmental 4412 Tryon Road Specialties Inc. Raleigh, NC 27606-4218

site Web Site: http:l /w~w.electrotechsystems. cam E-mail: [emailprotected] Web Site: http:/ lwww.egc.com

Phone: 9 19-829-9300 E-mail: [emailprotected] Fax: 919-833-9476 Web Site: http: l/www.esionline.com

ESPEC Corp. 425 Gordon Industrial Court Phone: 616-878-0270 E-mail: [emailprotected] Grand Rapids, Fax: 616-878-0280 Web Site: MI 49509-9506 Toll Free 800-537-7320http: lIwww.espec.com Forma P.O. Box 649 Scientific Inc. Marietta, OH 45750

Phone: 740-373-4763 Fax: 740-373-6770

Gilson Co. Inc.

P.O. Box677 Worthington, OH 43085

Phone: 800-444-1 508 E-mail: [emailprotected] Fax: 740-548-5314 Web Site: http: llwww.globalgilson.com

Hotpack

10940 Dutton Road Philadelphia, PA 19 154

Phone: 215-824-1700 Fax: 215-637-0519

E-mail: [emailprotected] Web Site: http: l1www.hotpack.com

Lunaire

2121 Reach Road Williamsport, PA 17701

Phone: 570-326-1770 Fax: 570-326-7304

E-mail: marketing~lunaire.com Web Site:

E-mail: fmarketing~forma.com Web Site: http: l1www.forma.com

Lunaire 1719 Route East, 10Suite Environmental 302, Parsippany, NJ 07054

Phone: 800-586-2473 Fax: 973-540-0367

Luwa Lepco Inc.

Phone: 713-461-1131 Fax: 71 3-464-1 148

E-mail: [emailprotected] Web Site: http: l1www.luwalepco.com

Powers P.O. 268 Box Scientific Pipersville, 18947 PA Refrigeration & Incubators Inc.

Phone: 215-230-7100 Fax: 21 5-230-7200

E-mail: mioff62voicenet.com Web Site: http: l1www.powersscientific.com

RKI lClean Room Services

P.O. Box16372 Rochester, NY 14616-0372

Phone: 800-447-4754 Fax: 776-621-2778

E-mail: [emailprotected] Web Site: http: I lwww.cleanroomservices.com

Rotronic Instruments Inc.

160 E Main Street Huntington, N Y 11743

Phone: 516-427-3898 Fax: 5 16-427-3902

E-mail: sales@rotronic-usa,com Web Site: httD:l1www.rotronic-usa.com

Sanyo Gallenkamp

900 N Arlington Heights Road, Suite 320 Itasca, IL 60143

Phone: 630-875-3543 Fax: 630-775-0427

E-mail: tharding~compuserve.com

1750 Stebbins Drive Houston, TX 77043-2807

Phonelfax

E-mail /web site

Scientific Orville 70Drive Industries Inc. Bohemia, NY 11716

Phone: 516-567-4700 Fax: 516-567-5896

E-mail: [emailprotected] Web Site: http: //wwwscind.com

Servicor

830 Bransten Road San Carlos, CA 94070

Phone: 650-591-0900 Fax: 650-591-3121

E-mail: [emailprotected] Web Site: http: //www.cleanroom.com

Spectrum Quality Products

14422 S San Pedro Street Gardena, CA 90248

Phone: 310-516-8000 Fax: 310-516-9843

E-mail: [emailprotected] Web Site: http: /lwww.spectr~chemica1. cam -

Sun Electronic 1900 Shepherds Drive Systems Titusville, FL 32780

Phone: 407-383-9400 Fax: 407-383-9412

E-mail: [emailprotected]

Surface 24040 Camino Avion, del Measurement Monarch Beach, CA 92629 Systems

Phone: 949-495- 1897 E-mail: [emailprotected] Web Site: Fax: 949-495-7795 http: //www.smsna.com

AddressName

Terra Universal

700 N Harbor Blvd. Anaheim, CA 92805

Thomas 99 High Hill Scientific Swedesboro, VTI Corp

Road

NJ 08085

7650 W 26th Avenue Hialeah, FL 33016

W R 1310 Goshen Pkwy. Scientific West Chester, PA 19380 Products

Phone: 714-526-0100 E-mail: info~terrauni.com Fax: 7 14-992-2 179 Web Site: http:/ /www.terrauni.com Phone: 856-467-2000 Fax: 800-345-5232; 856-467-3087

E-mail: [emailprotected] Web Site: http:/ /www.thomassci.com

Phone: 305-828-4700 Fax: 305-828-0299

E-mail: [emailprotected] Web Site: http: / /www.vticorp.com

Phone: 800-932-5000

Web Site: http: / /www.vwrsp.com

the storage conditions, the package description, and several other factors. The stability protocol is a detailed plan for the conduct of stability studies to generate stability data and their analysis usingstatistical principles whereapplicable, support of retest of drug substance, and for the expiration dating of drug products at labeled storage conditions. For drug products, acceptable labeled storage conditions include controlled room temperature, refrigerator temperature, or freezer temperature. In addition, stability protocols should be used to support an extension of the retest period or the expiration dating period through annual reports. A stability protocol for product registrations should also include a commitment that labeled storage temperature stability data for the first three post-production batches willbe generated to justify confirmation and extension of the expiration dating period. The commitment also includes monitoring of the stability of annual batches after the first three production batches. The commitment actually constitutes an agreement with the Food and Drug Administration

Suppliers and ~anufacturersof Photostability Test Chambers Phone/fax

E-mail /web site

Name

Address

Caron Products & Services, Inc.

PO Box 715, Products Lane Phone: 740-373-6809 E-mail: caron@fro~et.net Fax: 740-374-3760

Web Site: http: I 1~w.caronproducts.com

Environmental 510 E Washington Street Chagrin Falls, Growth OH ~ 0 2 2 - ~ 4 8 Chambers

Phone: 800-321-6854 Fax: 440-247-8710

E-mail: [emailprotected] Web Site: http: // ~ . e g c , c o m

Environmental 4412 Tryon Road Specialties Inc. Raleigh, NC 27606-42 18

Phone: 919-829-9300 Fax: 919-833-9476

E-mail: [emailprotected] Web Site: http: //m.esionline.com

Phone: 570-326-I770 Fax: 570-326-7304

E-mail: [emailprotected] Web Site: htttxl /www.lunaire.com

P.O. Box268 Powers Pipersville, PA 18947 Scientific Refrigeration & Incubators Inc.

Phone: 215-230-7100 Fax: 21 5-230-7200

E-mail: [emailprotected] Web Site: http: I /~w.powersscientific.com

550-29 E Main Street Branford, CT 06450

Phone: 203-483-5810 Fax: 203-481-8589

Web Site: http: I l ~ . ~ t r a ~ o l e t l a m p s . c o m

Lunaire

Southern New England Ultraviolet

Marietta, Ohio 45750

2121 Reach Road Williamsport, PA 1770 1

indicating that the plan outlined in the product registration document will bestrictly complied with. The commitment also requires submission of stability data at periodic intervals and/or annually as specified in the application. Finally, the commitment mandates that any batches thatare found outside of the specifications approved with the product will bewithdrawn from the market. If the deviation from approved speci~cationsis observedto be a single occurrenceand is not expected to affect the safety and efficacy of the product, regulatory guidance allowsnotification to and discussion with the approving regulatory division with detailed justification for distribution of the product.

tandard operating procedures (SOPS),as the name suggests, are written procedures describing operational procedures. Current good manufacturing practices (cGM andate thatevery organization develops a set ofSOPs to describe their operations. is almost impossible to develop anystability management function without having a comprehensive set of SOPS. Standard operating procedures cover all as ects of operations. In the United States, numerous citations (Form 483 Inspection are issuedeveryyear to pharmaceutical companies to point out operational deficiencies,whichmust be addressed. Frequently, the regulatory citations are the cause of recommendations by FDA field inspectors to withhold approval of product registrations. These citations often are issued as a result of Pre-Approval Inspections. Many of those citations deal with two key principles, one with lack

of standard operating procedure, the other with inconsistencies with practice and dealing with differences between practice and what is conch more severe than lack of SOPs. We do not wish to condone if one wereto choose to err, itis better to not have an S follow it. Whilespecific SOPs to address specific should be written in general flexibleform without including rigid that may make them too restrictive and require frequent revision. be periodically reviewed and revised as needed. S not intended to cover all aspects of operating procedure appliowever, a list of significant SOPs that must be developed can be constructed. The following list isbased on numerous form 483 citations that have been issued during the past few years. It should be used as an instructive list that is more likely to result in the issuance of regulatory citations, if not followed. The list is not intended to be an exhaustive but merely covers significant issues that must be addressed in SOPs. There should bea written procedure establishing a time limitfor samples, after removal from the stability cabinets, awaiting laboratory analysis. The written program must list the acceptable procedure and practice. It should document a rationale for acceptance of data generated. An assurance of the integrity of the sample should be supported by the procedure. uld address how “out of specification” results are handled why retesting is performed must be addressed. This ap in-house as well as outside contract laboratories. How an investigation is carried out? Who signs the investigation report? Under what circumresults classified as “inconclusive” immediate retesting is re follow-up investi ations done to determine the reason o isresponsible for performing and evaluating such henever an investigation of an “out of specification” result leads to a lab-related issue, the procedure should call for extending the investigation to determine if other lots or other product analyses could have been likewise affected, by the equipment or analyst. There should be a procedure to evaluate stability trends. This pertains to systematic downward or upward trends inany particular test result. For example, it is not uncommon to see a downward shift in in-vitro dissolution test results. There should be a written procedure establishing the steps for inactivating (closing) stability projects in the stability system scientific software, whenever applicable. The procedure should address the criteria for closing or inactivating projects and a detailed explanation for documenting the reaso~sfor these activities.Stability projects or stability data in the software system should neverbe deleted. Any changesfor data or projects should be made in the form of amen~mentsthrough clear documentation stating the reason and justification for doing so. A procedure shouldexist,whichoutlineshow the periodic reviewof temperature/relative humidity charts from stability rooms and cabinets or continuous monitoring and surveillance of stability rooms and cabinets

is conducted. It should address the procedure for handling normal reports as well as how the discrepancies, deficiencies, deviations, or malfunctions in the system are detected. Who is responsiblefor investigating deviations?There should be a procedure for documenting that stability samples arenot adversely affected. The procedure should describe how corrective actions are undertaken. Stability procedures dealing withanalytical me tho do lo^ should address how and whensignificantchanges in analytical test results are investigated. For example, and adequate investigation needs to be undertaken when an additional peak shows on a chromatogram compared to previous results. The investigation should determine and document the cause of the additional peak or other significant change. Procedures for preventive maintenance, including calibration of stability rooms, cabinets, and laboratory equipment, instrumentation, balances, monitoring equipment, etc. should list the frequency of testing at periodic intervals and documentation of results. ords for calibration should be maintained for specified periods. In add n, the calibration tests, which are performed by a vendor, should be described in a written procedure, contract, or report. ocumentation of label preparation, review, and approval should be addressed. It should address the retention of label specimens with batch records. Further details regarding the identity of the dates, the persons performing the operations, the equipment used, and any control procedures employed, and yield determination should be included in the procedures. A procedure for the operation and control of the computer software used for input and output of stability data, including verificationof the data, should be developed. The procedure should address whois authorized to flag erroneous data and enter corrected data whenever errors are discovered. The procedure should clearly state that data should never be deleted. There should be a procedure available to ensure proper handling of stability samples whosedue date falls on Saturdays, Sundays, and known holidays. The procedure should clearly delineate how and when the samples are actually pulled.In general, the practice of pulling samplesafter their actual test date rather than before the test date can be defended better on scientific grounds.

There should be a written procedure that describes a training program for laboratory personnel to assure that the employees are trained in and remain familiar with laboratory control procedures and current good manufacturing practice regulations. In addition, the procedures and criteria used to qualify or certify employees in specific analytical methods should be described in the written programs. Regulations require that “each person engagedin the manufacture, processing, packing, or holding of a drug product should have education, training, and experience, or any combination thereof, to enable that person to perform the assigned’ functions. Training shall be in the particular operations that the employees performs.” Training must be conducted by qualified individuals on a con ti nu in^

basis and with sufficient frequency to assure that employees remain familiar with aintenance of training records is an equally important proyers are able to provide adequate training but fail to maintain appropriate records of training.

uring the discovery phase,stability evaluation of almost every batch of drug substance used in the conduct of preclinical safety testing and/or for clinical trials is carried out. The purpose of this is to establish a database and a history of critical stability characteristics of drug substance. ese data provide a foundation for justifying a retest period for drug substance. cause of the paucity of data, initially a shorter retest period is commonly employed.A longer retest period can be justified as more stability information is generated. The data is essentialfor the establishment of stability trends and also for setting of specifications for the drug substance. These data serve as the basis of supporting information with the registration applications. A s the product continues through further clinical testing and reaches the marketable stage, stability testing for at least three batches of drug substance at normal storage and at accelerated conditions must be available with the product registration application. Generally a minimum of twelve months long-term stability data is sufficient for filing purposes. Selection of three batches provides a means for evaluation of statistical variation in the process developmentand demonstrates the consistency of the results. These three batches must be made on a pilot plant scale. They must usethe same synthetic or biological process for manufacture that is intended for manufacture on a large-scale commercialproduction. The synthetic procedure must represent the evolutionary development from earlypreclinical and clinical testing stages. The packaging for drug substance for long-term stability testing must simulate the intended drug substance storage. It is very common to employ smaller representative samples of the same material composition and to simulate a comparable drug-to-surface-area ratio as would be expected on large-scale storage. The overall quality of the three-batches must be comparable to or better than the quality of the batches used in early stages of development. If the three batches of drug substances included in the product registration application do not represent the full-scale manufacture, then the first three production batches of drug substance post approval must be placed on the same long-term stability evaluation protocol as was used earlier. imilarly, during the early clinical testing (IND) stage, stability evaluation of several batches of drug product used inthe conduct of clinical testing is carried out. The purpose of the data generation is to establish a stability database and identify critical stability characteristics of the drug product. The goal is to identify the expiration dating period for clinical formulations. During product development stages, formulation development continues through further refinement in terms of strengths, size and shape where applicable, and scale-up to support larger clinical trials. Since these stability data may be a multitude of packages and may represent minor changes in the formulations, they serve as the supportive data for product registration purposes. The primary stability data for submission withthe product registration application must be generated on at least three batches of drug product packaged in

the containerlclosure system intended for marketing. Primary stability studies are intended to show that drug product stored in the proposed container/closure for marketing will remain within specifications if stored under storage conditions that support the proposed shelf life. Guidelines suggest the submission of at least 12 months labeled storage temperature data at the filing time. In addition, at least 6 months accelerated stability data should be included with the a~plication.Two of the three batches of drug product should be at pilot scale. not much of a problem for products indicated for chronic therapeutic uses where clinical trials require large-scale manufacture of clinical supplies. The third batch may be from a smaller lot (for example, 25,000to 50,000 tablets or capsules in case of solid dosage forms). Preferably these three batches of drug pr manufactured from three different batches of drug substance. for the manufacture of product must be from the synthetic process intended for commercial production. The process usedfor the manufacture of drug product must be representative of what is intended for large-scale manufacture. It should provide acceptable continuity with batches manufactured during early development stages. The overall quality of the product should be the same quality intended for marketing, and it should meet all the specifications included in the registration document.

Selection of stability storage test conditions for conducting long-term stability studies is based on an analysis on the effects of climatic conditions existing in the regionwhere the drug product distribution isexpected to occur. For multinational pharmaceutical companies, stability testing can becomeunwieldy, if a separate study were to be conducted for each country. Fortunately, the IC guidelineshave made the task easier by adopting the climatic zonesconcept. The concept of mean kinetictemperature in any regionof the world is nicely defined in terms of climatic zones(Grimm, 1985,1986; aynes, 1971). Based on the analysis, the world is divided into four climatic zones, to which the individual c assigned. The four climatic zones are defined as I. Temperate climate, 11. and ~editerraneanclimates, III. Hot,dry climate, and IV. Hot, humid climate. The climate zone I1 covers the European Union, Japan, and the United State. The ICH Guidelines and the recent FDA draft guidancerecommend the long-term stability storage at 25°C f2°C 60% RH f5% for most stable products and for manysoliddosage forms. The length of the stability studies and the long-term storage condition should be sufficient to cover shipment, storage, and subsequentuse(forexample, reconstitution or dilution as recommended in the label). For products whose long-term storage can b ed at 25°C f2"C 60% f5%, accelerated conditions at 40°C f2°C 75% f5% isrecommended in the guidelines. Long-term stability study generally should be scheduled for up to 60 months at 25°C f2°C 60% RH st 5%. A minimum of 12 months stability data at the long-term stability storage (25°C f2°C 60% RH f5% in this case) should be submitted with the registration application. A minimum of 6 months data at accelerated condition (40°C f 2 " C 75% RH f5%) should be available along with the long-term stability data. The regulatory guidance suggests, "when significant

change occurs due to accelerated testing, additional testing at anintermediate conf 5 % ) should be conducted” (F dition (for example,30°C f 2”C/60% nt at the accelerated condition is defined as 19988). ~ i g n ~ c a change 1. A 5 percent potency loss from the initial assay value of a batch 2.Anyspecified degradant ex ding its specificationlimit 3. The product exceeding its exceeding the specification limits for 12 capsulesor 12 tablets eet specificationfor appearance and physical properties (e.g., color, phase separation, resuspendibility, delivery per actuation, caking, hardness)

f5%, a minimum henever significant change occur at 40°C f2°C75% ths’ data from an ongoing l-year study at 30°C 60% should be included with the initial registration application with the provision that same significant change criteria apply. The draft guidance (FDA, 1998b) includes r e c o ~ e n d a t i o n sregarding continued occurrence of significant changeduring 12 months’ exposure at 30°C f 2°C f5% as well as for several other situations. or products, which are stable at refrigerated conditions, long-term stability storage at 5°C f3°C should be conducted. onitoring of relative humidity but not its control is recommended at the refrigerated storage temperature. Generally, accelerated conditions at 15°C higherthan the long-term stability storage conditions can be used as the accelerated conditions. Following the principle, accelerated conf5% are appropriate for refrigerated storage ditions of 25°C f 2 ” C 60% product. Likewise,when the duct isextremelysensitive to temperature and requires freezerstorage, long-term storage at -15°Cf5°C with accelerated testing erated temperature condition of 5°C f 3°C is recommended. raft guidance includes several specialsituations. For example, for liquids in glass bottles, vials, or sealed glass ampoules, which provide an im~ermeablebarrier to water loss, the following factors should be taken into consideration. Accelerated condition, 40°C and ambient humidity isan acceptable alternative to 40°C f2°C 75yo RH f 5%. Intermediate condition, 30°C and ambient humidity is an acceptable alternative to 30°C f 2°C 60% R H f5%. Long-term condition, 25°C and ambient humidity is an acceptable alternative to 25°C f2°C 60% situation is that it requires stability One of the complicationswith these special rooms and/or cabinets at several identicalconditions with minor differences. these can not be justified economically, as they add to the total expense not only for acquisition but also for continuous monitoring, maintenance, and record keeping. Asan aid to the readers interested in pursuing the establishment of stability rooms and/or cabinets, Table l provides a list of equipment suppliersand/or manufacturers in alphabetical order. The list isnot intended to be exhaustive. The reader is cautioned that inclusion in this list doesnot constitute an endorsement of these suppliers in any way. These are provided as an aid only. Likewise exclusion of other important suppliers in the list does not mean that they are not reliable and thus

very effort has been made to check the accuracy of the list ase convey any errors to the author.

ecially solid dosageforms (tablets, capsules), oral powders, e not affected by short-term exposure to va ' untered during ship~ingand distribution. r example ointments, creams, suspensi Is, and sup~ositoriesmaybeadverselyaffected by rature fluctuations. They mayundergo phase separation, n, changes in viscosity, creaming, sedimentation, and so on. These typesof drug products should be tested under cycling conditions to simulate shipping and distribution conditions. These ed on the packaged product during stress testing of ting parameters for those studies should i also physical changes and homogeneity. several recommendations for thermal *

studies. ducts that may be exposed to above-freezing temperature variations, a erature cycling study consisting of three cycles of 2 days orrefrigerated ge ( 2 4 ° C ) followed by 2 days under accelerated conditions (40°C) should be performed. oducts that may be exposed to subfreezing temperat~revariations, a erature cycling study consist in^ of three cycles of2 days at freezer temrature (- 10 to -20°C) followed by 2 days under accelerated conditions "C) is recommended. en drug products, the recommended cycling study should include an evaluation of effects due to thawing under a hot water bath or in a microwave unless the drug in the product is known to degrade at high temperatures. or inhalation aerosols, the recommended cyclingcondition consisting of three to four cycles of 6 h per day, between subfreezing te -20°C) and accelerated conditions (40°C ~5%-85% to 6 weeks should be followed. c drug product characteristics may require an alternative cycling condition study as long as it is scientifically justified.

ite Guideline as well as the draft guidance (IC

,1998b) state that light testing should be an integral part of stress testing. u r i n ~the develop~entphase, the intrinsic photostability characteristics of new rug substances and products should be determined. The purpose of such study is to demonstrate that light exposure does not result in unacceptable change and to determinewhether precautionary measures in manufacturing, labelling, and pac~agingare needed to overcomechanges due to lightexposure. Generally, hotostability testing is carried out on a single batch of material selected.

hotostability information is generally requiredfor submission in registration tions for new molecular entities and drug products made from the entity. Guideline and the draft guidance recommend a systematic approach to photostability testing to cover studies such as Tests on the drug substance Tests on the exposed drug product outside of the immediate pack, and if necessary Tests on the drug product in the immediate pack, and if necessary, Tests on the drug product in the marketing pack. cision FlowChart for Photostability Testing of Drug Products included guidelinesuggests the extent of drug product testing by assessing whether acceptable change has occurred at the end of the light exposure testing. Acceptable change is defined as change within limits justified by the applicant. enerally, the photostability studies are carried out by exposing the samples to specific light sourcesas defined inthe guideline. The guideline offers twooptions for light sourcesfor testing. The first is“any light sourcethat is designedto produce an output similar to the 651ID65 emission standard such as an artificial daylight flucombining visible and ultraviolet (UV) outputs, xenon, or metal 65 is the internationally recognized standard for outdoor daylight. ivalent indoor indirect daylight standard.’’ For option 2, the same sample should be exposed to both a cool white fluorescent and a near-ultraviolet ent lamp meeting the criteria specified. lnce temperature changesoccurwhen the samples are exposed to light sources, protected samples (for example samples wrapped aluminum in foil) are also exposed side by side with the authentic sample so that any contribution due to thermal exposure can be determined in the dark control sample. For the drug substance, photostability testing is dividedinto two parts, forced stress testing and confirmatory testing. Generally, forced stress testingis done by exposing the drug alone and/or in solutions/suspensions to evaluate the photosensitivity, to validate the analytical method, and to elucidate degradation pathways. onf fir ma tory studies should then be undertaken to generate the information necessary for handling, packaging, and labeling. For confirmatory studies the guideline suggeststhe exposure of samples to light providing an overall illumination of not less than 1.2 million lux hours and an integrated near ultraviolet energy of not less than 200 watt hourdsquare meter to allow direct comparisons between the drug substance and the drug product. The guideline recommends the conducting of studies on drug products to be carried out in a sequential manner starting with testing the fully exposed product and then progressing as necessary to the product in the immediate pack and then in the marketing pack. Testing should continue until the results demonstrate that the drug product is adequately protected from exposure to light. Normally, only one batch of drug product istested during the developmentphase, and the photostability characteristics should be confirmed on a single batch of drug product selected. It should be noted that photostability testing is stress testing designed to determine the intrinsic photostability characteristics ofnew drug substances and drug products. No correlation has been developedor established that can equate the within-specificationresult to the labeled expiration dating period. Detailed

description of the presentation of samples, the analysis of samples, and the judgment of results is covered in the guideline and the draft guidance. cerning an actinometric procedure for monitoring exposure to a near-UV fluorescent lamp and other actinom~tricprocedures and light sources are included in the guidelines. The reader is also referred to several references on the subject (Drew et al., 1998a and 1998b; Tonnesen, 1996; Sager et al., 1998; Yoshioka et al., 1994). The draft guidance also provide an account of acceptable/unacceptable photostability change and photostability labeling considerations. As an aid to the reader interested in pursuing the e s t a b l i s ~ e n t of photostability testing, Table 2 provides a list of equipment suppliers and/or manufacturers in alphabetical order. The list is not intended to be exhaustive. is cautioned that inclusion in this list does not constitute an endorsement of these suppliers in any way. Theseare provided as anaid only. Likewise exclusion ofother important suppliers in the list does not mean that they are not reliable and thus should not be considered. Every effort has been made to check the accuracy of the list contents. Please convey any errors to the author. yste

The draft guidance recommends that long-term stability data should be generated for the drug product in all types of sizes of immediate containers and closures proposed for marketing, promotion, or bulk storage. Good science dictates that rigorous studies be undertaken to test any possible interaction between drug product and the immediate container and/or closure. This includes consideration of leachables or extractables from the container /closure into the drug product environment or surface. Similarly, absorption or adsorption of product co~ponentsinto the container/closure should be investigated for liquid or semisolid products. The studies must employ meaningfuland sensitive assay methodologyto detect this type -oral solutions, sterile solutions in small volume container systems, and S in large volume parenteral container/closure system should be tested in upright and inverted and/or on-the-side positions at both long-term and accelerated conditions to ensure that no adverse effect from any interactions is produced. A direct comparison of the data from upright and inverted or on-the-si~e~ o s i t i odemonstrates ~ whether any extractable and/ or absorption/ n has occurred or is likely to occur. gulatory guidelines do allow for conducting bracketed stability studies for intermediate containers when the drug product is packaged in the smallest and the largest containers, provided that all containers and closures share the same composition and design.

Testing frequencyfor both drug substance and drug product for product registration should be sufficient to establish stability characteristics. The ICH Guidelines and draft guidance require that testing should be every 3 months over the first year, every 6 months over the second year, and annually thereafter. For drug products packagedinidentical container /closures of differentsizes and/ or identical formulations of different strengths, matrixing and/or bracketing can be used with

valid justification. uring early evaluation phase, it is not uncommon to use more frequent, for example every 2 weeksduring the first 3 months. For accelerated stability conditions, testing is generally done for 6 months. The guidance recommends testing at 0, 2, 4, and 6 months so that at least four stability time-points are generated.

A critical component of stability testing is the development of stability indicating assay methodology. Analytical methodology should be developed and validated to assure that the method can accurately and quantitatively analyze the major component(@, impurities, degradation products, and other selected components in the drug substance and drug products. evelopment and validation of the method should ensure that the method is able to analyze the com~onent(s)in the presence of other components. Methodologydevelopmentshould address key analytical and statistical parameters, such as accuracy, precision, detection and quantitation limits, linearity, range of analytical limits, robustness, reproducibility, repeatability, ruggedness, and system suitability testing. The standard operating procedures must describe the procedure in sufficient detail, the steps necessary for performing the test, including the sample preparation, reference standards, preparation of reagents, instrumentation used, calibration, calculations, and so on to ensure the validity of data. A partial revalidation of assay methodology should be undertaken any time there is a significant change in the product formulation or in the process that may affect the validity of the original assay.

Drug products that require reconstitution prior to dispensing to the patient or the consumer require special consideration with regard to stability testing. Section CFR states that “testing of drug products for reconstitution at 211.166of21 the time of dispensing (as directed in the labeling) as well as after they are reconstituted’’shall be includedin the written program for stability testing. ~xamplesof reconstituted products includelyophilized products, powders or granules for reconstitution, dilution of active product with additives, admixture of drug product with other components, and so on. For reconstituted products, two distinct stability periods are in operation. The first period coversup to its expiry through long-term and accelerated stability testing prior to hesecondperiodconsistsofusually short-term stability after ability study after reconstitution as directed on the label should be conducted by appropriate sampling and testing to cover a period beyond that specified on the label. For example, if the reconstituted product were recommended to be used within l week under refrigeration, then an appropriate stability study could covertesting at periodic intervals up to 10 days or 2 weeksunder refrigeration. ddition, an accelerated condition stability of reconstituted product at 25°C 60% for 1 to 3 days may be appropriate.

The draft guidance provides specific recommendations for evaluation of stability attributes specific to many dosage forms. In general, assay, appearance, and degra-

dation products should be evaluated for all dosageforms. It should be noted that not all tests need to be done for all dosage forms, nor it is intended to be an exhaustive list. Selection of attributes for stability evaluation should be made on the basis of stability behavior of new drug product observed during development process. Finally, it is not expected that every listed test be performed at every time point. Not all the dosage forms included in the draft guidance are discussedbelow. For the complete list, the reader should refer to the draft guidance. The draft guidance recommends that Tablets should be evaluated for appearance, color, odor, assay, degradation products, dissolution, moisture, and friability. ard gelatin capsules should be evaluated for appearance (including brittleness), color, odor of contents, assay, degradation products, dissolution, moisture, and microbial limits, Soft gelatin capsules should be tested for appearance color, odor of contents, assay, degradation products, dissolution, microbial limits, and pellicle formation. In addition, precipitation and cloudi medium should be examined. Emulsions should be evaluated for appeara includingphase separation, color, odor, assay, degradation products viscosity,microbiallimits, preservative content, and meansize an ibution of dispersedphase globules. Oral solutions shouldbetested for appearance (including precipitate formation, clarity), color, odor, assay, degradation products, pH, microbial limits, and preservative content. Oral suspensionsshouldbe aluated for appearance, color, odor, assay, degradation products, p microbial limits, preservative content, redispersibility, rheological properties, and mean size and distribution of particles. Oral powders for reconstitution should be evaluated for appearance, color, moisture, and reconstitution time. After reconstitution, testing for oral solutions or suspensions should be applied. etered-dose inhalations and nasal sprays should be evaluated for appearance (including content, container, valve, and its components), color, taste, assay, degradation products, assay for cosolvents, dosecontent uniformity, labeled number of medicated actuations per container, aerodynamic particle size distribution, microscopic evaluation, water content, leak rate, microbial limits, valve delivery(shot weight), and extractables/leachables from plastic and elastomeric components. Samples should be stored upright and inverted or on the side. In~alationsolutions should be testedfor appearance, color, assay, degradation products, pH, sterility, particulate matter, preservative and antioxidant content (if present), net contents, weight loss, and extractables/leachables from plastic, elastomeric, and other packaging components. Topical, ophthalmic, and otic preparations include ointments, creams, lotions, pastes, gels, solutions, and nonmetered aerosols for application to the skin. Topical preparations should be evaluated for appearance, clarity, color, homogeneity, odor, pH, resuspendibility (for lotions), consistency,

viscosity, particle size distribution (forsuspensions), assay, degradation products, preservatives, and antioxidant content, microbial limits/ sterility, and weight loss. Small volume parenterals (SVPs). ncluded in this category is a wide range of injection products such as drug injection, drug for injection, drug injectable suspension, drug for injectedsuspension, and drug injectable emulsion. Stability evaluation of drug injection should include appearance, color, rvative content (if present), degradation products, particulate sterility, and pyrogenicity. Drug for injection should include f appearance, clarity, color, reconstitution time, and residual moisture content. After reconstitution, stability should be monitored for attributes listed in drug injection. Large volume parenterals (LVPs) should be evaluated for appearance, color, rvative content (if present), degradation products, particulate sterility, pyrogenicity, clarity and volume.

A lot of products are marketed in more than one type of product configuration. For example, tablets are packaged in multiple sizes, which may include a very small size for promotional use, several intermediate sizes for pharmacy or patient use, and large sizes for institutional usage. A similar situation may exist with liquid products, semisolids, and others. Other examples may include tablets in several strengths packaged in identical container/closure systems. The useof bracketing in those situations to reduce the full stability evaluation is important in stability studies. A detailed account of bracketing designs and other complex designs are included in recent publications (Fairweather et al., 1995; Nordbrock, 1992). Bracketing maybe definedas the design of a stability testing scheduleso that at any timepoint only the stability samples on the extremes, for example, of container size andlor dosage strengths, are tested. The key assumption for the bracketing design is that the stability of the intermediate size samples is represented by that of the extremes. A bracketing design may be used for primary stability studies to be included in the registration application, stability batches as committed after approval of the product, and annual batches or batches included in supplemental d generally not be used for clinical formulations racketing design S during the development phas otocols for bracketing desi by the FDA prior to initiati primary stability studies Draft guidance providesa detailed discussion of brack design isapplicable to most dosage forms, including immediate release solid dosage forms and modified release oral solids, liquids, semisolids, and injectable products. It may not be readily applicable to certain types of dosage forms, such as metered dose inhalers, dry powder inhalers, and transder a1 products, unless convincing scientific arguments can be justified. The Food and rug Administration encourages and welcomes consultation on the subject. The simplest situation for bracketing design is wherethe same drug product is packaged in different sizelfill containers of the same composition. Where the range of fill/size of the same strength drug product is to be evaluated, bracketing may be applicable if the compositions of the containers and closures are identical

throughout the range. Bracketing design isapplicable without further justi~cation, where the range of size or fill varies while other factors are constant. design is also applicable, where both the container size and the fill var good justification is included and supported by data. ~ustificationsh discussion of the surface-area-to-volume ratio, the dead-space-to-volume ratio, the container wall thickness, and the geometry of the closures. S long as the intermediate sizes are adequately bracketed by extremesizes, the bracketing design should be justifiable. racketing design is suitable and applicable to identical or closely related formulations packaged in different size identical container /closure systems. Examples includetablets compressed to different weightsfrom the S or powder mix. Similar examples includea range of different size different plug weightsfrom the same granulation. is involved, such as different amounts of active in e amount of each excipient or the total weight of the dosage form remains constant, bracketing design plicable unless justified by supportive stability data from development eting designsare generally not applicable ~ h e n ~ vmultiple er variables exist, for example significantlydifferent formulation strengths made from different excipients, colorants, flavors, and so on. In selected cases where multiple variables are involved matrixing may be considered.

es definematrixing as the statistical design of a stability schedule so that only a fraction of the total number of stability samples are tested at any specified sampling point. At a subsequent sampling point, different sets of samples of the total number are tested. The design assumesthat the stability of the samples tested represents the stability of all samples. The reduced stability testing in the matrixing design offers an excellent alternative for monitoring stability where multiple factors for the same product, such as different stre ths of the same formulation, different batches, differentsizes of packages, a differentfills are involved. Other factors, such as different container and closure systems, different strengths of closelyrelated formulations, different orientations of containers during storage (upright and inverted or on the side), multipledrug subst ce manufacturing sites, multiple drug product manufacturing sites, may be matrix with appropriate scientificjustification. Supportive stability data from the development ismost crucial for justifying complex matrix designs. For example, for primary for~ulationsthat show excellent stability to moisture, oxidative exposure, and light exposure are less likely to be affected by differences in package configurations versus those that show marginal stability against these variables. It may be poss for different packages, such as glass bottles, blister pac polyethylene bottles for such a formulation with excelle scientific grounds, it may bemore ~ifficult justify to matrixing sig~ficantlydifferent formulations with different excipients, unlessthe drug substance bas been found to be highly stable and almost unaffected by most environmental variables. The draft guidanceprovides a detaileddiscussion of matrixing design. atrixing design is applicable to most dosage forms, including immediate release solid dosage forms, modified release oral solids, liquids, semisolids, and inj~ctable

may not be readily applicable to certain types of dosage forms such as e inhalers, dry powder inhalers, and transdermal products unless convincing scientific arguments can be made. The regulatory guidelines require that atches be tested initially and at the end of the long-term testing, rug Administration encourages and welcomes consultation on Factors affecting the applicability of matrixing design for primary stability studies are the stability of drug substance, drug product formulations during development stages, and the variability in the stability data, other than analytical method variation. Stability data variability for matrixing design purposes refers to batch-to-batch variability, strength-to-strength, size-to-size, and across different variables, such as batch versus strength, strength versus size. In general, the better the stability behavior of the drug substance and formulations, the better the applicability of matrixing when accompanied by less variability in the data. Likewise, the opposite is true when the matrixing design should not be applied in situations with poor stability behavior and larger variability. variability means matrixing may be applied witha historical data from primary stability determineswhether may be applied to post-approval commitment batches or supplemental batches. The draft guidancesuggests that for product registration applications, a matrixing design should always include the initial and final test points, as well as two additional points through the first 12 months. All samples including those that are scheduled to be tested should be placed on the stability program. The protocol should be followedas planned without deviation once the study has begun. The size of the matrixing design is expressedas a fraction of the total number of the sample to be tested in the full corresponding stability protocol. The size depends on the amount and quality of supportive data as stated earlier and the number of variables to be matrixed. For example, for a product with excellent stability available in three batches, with three different strengths, and packaged in three packageshill sizes, the number of factors to be tested in a full stability protocol is 3 x 3 x 3 x 27. For such a product, the size of the matrixing design could be as small as half the original full stability protocol. Thus fractional 3means that only halfas many samplesas the full stability protocol will be tested in the matrixing design. Oneof the key elements of matrixing design is that it should be a well-balanced design. If possible, an estimate of the probability that stability outcomes from the matrixed study would be the same for a given factor or across factors should be provided with the protocol. Several different fractional matrixing designs are included in the draft ,1998b). A detailed account of matrixing designs and other complex analysis is provided in recent articles (Fairweather et al., 1995; Nordbrock, 1992).

The presentation and evaluation of the stability information should be based on a systematic approach. Evaluation of stability should cover all chemical assay data and,as necessary,physical,chemical,biological, and microbiological quality characteristics, including particular properties of the dosage form (for example, in-vitro dissolution rate for oral solid dose forms).

The goal in the design of the stability study is to help establish a shelf-life and label storage conditions that can be applied to all future batches of the dosage form manufactured and packaged under identical conditions. This is accomplished by testing a minimum of three batches of the drug product under a variety of stability conditions. Generally, the better the stability of the product, the better is the confidence that a future batch will remain within acceptable specification until the expiration date. Likewise, the greater the degree of variability of individual batches, the greater will it influence the confidence that a future batch will remain within acceptable specification until the expiration date. Generally, an expiration period for drug products and a retest period for drug substances should be determined on the basis of statistical analysis of the observed long-term stability data. The draft guidance allows limited extrapolation beyond the observed range to extend the expiration dating or retest period, provided it is supported by the statistical analysis of the real-time data, satisfactory accelerated data, and other supportive data. For the purpose of the d e t e r ~ n a t i o nof expiration date, the initial value for the batch is critical. The initial value is interms of a percentage of the label rather than a percentage of the initial average of the results. The timeduring which a batch can be expected to remain within speci~cationsdepends not only on the rate of physical, chemical,and microbiological changesbut also on the initial average value for the batch. The expiration dating period for an individual batch is based on the observed pattern of change in the quantitative attributes (for example assay, degradation products) under study and its precision. The key is the quantitati~enature of the attribute. commonly acceptablestatistical approach for analyzing an attribute that is expected to decrease with time isto determine the time at which the 95% one-sided lower confidence limit (also known as the 95% lower confidence bound) intersects the lower specification limit. For most degradation pathways, since the overall observed degradation is less than 10% one cannot distinguish between a zero-order rate of degradation and a higher-order rate of degradation. Consequently, the rate of degradation can safely be assumed to be linear. For drug products, usually the 95% lower confidencebound at 90% of labeled claim is assumed. For the appearance of degradation products, which are expected to increase with time, the 95% upper limit for the mean should be used. en the 95% one-sided lower limit for assay and the 95% one-sided upper limit for degradation products are used, both the average assay values and the degra95% confidence. dation product values are likely to remain within specifications with The expiration dating period using the point at which the fitted least-squares line intersects the specification limit should not be used. It is as likely to provide an overestimate of the expiration dating period as it is likely to underestimate it in which case the batch average can be expected to remain within specifications at expiration if the fitted least-squares line is used with a confidence level of only 50%. Ifanalysisshows that the batch-to-batch variability is small, that is, the relationship between assay or degradation products and time is essentiallythe same from batch to batch, the stability data should be combined into one overallestimate. Combining the stability data from batches should be based on batch similarity. The

similarity of estimated curves among the batches tested should beassessedby applying statistical tests of equality of the slopes and of zero time intercepts. A detailed account of these principles is covered in the draft guidance. A useful concept is the release limits of drug products. The release limits of dosage forms are defined as the bounds on the potency at which an individual lot can be released for marketing which willensure that itremains within registered limits throughout its shelf life. A statistically based method is used for calculating release limits for any type of dosage form and any parameter for which the rate of change with time ispredictably uniform and linear. hen the mean refease assay result for a specific product lot is at or within the calculated release limit bounds, assurance is provided at the specified confidence levelthat the average assay results obtained at any subsequent time within the expiry dating period will remain within registered specifications ( Often the data may dation and so little variability that it is apparent from looking at the datathat the requestedshelflife can easilybe supported. Under the circumstances, it is normally unnecessary to go through the formal statistical analysis, one merelyprovides a full justification for the omission. ny evaluation should consider not only the assay but also the levels ofdegradation products and appropriate attributes. Where appropriate, attention should be paid to reviewing the adequacy of the mass balance, different stability behavior, and degradation performance.

evelopment of a stability information management system is the pharmaceutical industry requires careful con~iderationof organizational structure, its needs, and the desire to integrate many factors involved in stability data collection, storage, retrieval, analysis, computation, validation, and so on. The easy availability of personal computers and of access to the internet have made the development of such much closerto reality than what would have been possible just a few years r to the explosion of personal computers, only mainframe or minicomputer ere available for information management systems. Theserequired a lot of resources for the development of stability information systems operating systems and a whole host of computational languages expansion has made the development of specialized systems suc ecause of the unique nature and the limited market, only a very few commercially available systems existfor managing stability information; they are specifically designed to serve the needs of the pha~aceuticalindustry. rom regulatory perspectives, the management of a stability system or other laboratory data acquisitions syst requires careful consideration of the whole proA guidance related to the subject provides some cess of collection of the data. Th insight. ‘‘It isimportant, for computerized and non-computerized systemsto define the universe of data that will be collected, the procedures to collect, and the means to verify its accuracy. Equally important arethe procedures to audit data and the program and the process for correcting errors. Several issues mustbe addressed when evaluating computerized laboratory systems. These include data collection, pro-

cessing, data integrity, and security. Guidance on security and authenticity issues for computerized systems covers the following items: rovision must be made so that only authorized individuals can make data entries. Data entries may not be deleted. The data must be made as tamperproof as possible. he standard operating procedures must describe the procedures for ensuring the validity of the data. One basicaspect of validation of laboratory computerized acquisition requires a comparison of data from the laboratory system(s) withthe same data electronically processed through the system and emanating on a printer. riodic data comparisons would be sufficient only when such comparisons have en made over a sufto assure that the computerized system produces consistent ested in pursuing the ac~uisitionor development of stability information systems, Table 3 provides a list of developers of com~uterized systems and laboratory information mana~ementsystems in alphabetical nly three of the suppliers from the list are known to offer stability information systems (Lycoming Analytical, Metrics, and ScienTek Software). The list is not intended to be exhaustive. The reader is cautioned that inclusion in this list does not constitute an endorsement of these suppliers in any way. Theseare provided as an aid only. Likewise exclusionof other important suppliers from the list does not mean that they are not reliable and thus should not be c sidered. Every effort has been made to check the accuracy of the list contents. leaseconveyany errors to the author. A well-designed stability system: should provide the flexibility in customizing the system to suit the organization’s needs; should be user friendly; should follow logical operational procedure; and should be easyto validate. It should reduce clerical needs; should reduce turnaround time; should provide dependable storage of data; and should reduce errors. The system should reduce costs and improve capabilities; should improve efficiency; should provideeasy retrieval; should serve global needs; and should allow extensive search capability. The system should allow inclusion of attributes, that the organization may want or re~uireto investigate specific questions or for ma the i n f o ~ a t i o n , in a searchable field. For example, ifwe are interested in ning all stability batches of a product containing a specific batch of active ingredient? or all drug S in a specificpackage,unless the information isincluded in a ,it may not be possible to retrieve the information. The stability system should offer the following capabilities for the management of stability operations. Not all the capabilities may be needed or useful for an or~anization, ~ser-friendlysystem nitiation of stability of new batch ~tabilityprotocol modification activation of stability try, verification, and approval of data

Suppliers and ion Managemen

ers of Stability System Software

and Laboratory

Artel Inc.

25 Bradley Dr Westbrook, ME 04092

Phone: 207-854-0860 Fax: 207-854-0867

E-mail: [emailprotected] Web Site: http: l1m.artel-usa.

Beckman Coulter

90 Boroline Rd Allendale, NJ 07401

Phone: 201-818-8900 Fax: 201-818-9740

E-mail: 1aosal~s~bec~anlao.com Web Site:

Compex USA Inc.

2625 Butterfield Rd Suite 307E Oak Brook, IL 60523

Phone: 630-368-0905 Fax: 630-368-1086

Web Site: http: l1www.compex.be

Lycoming 2687 Euclid Avenue Analytical Duboistown, PA 17702 Laboratories (CF Technical Development, Inc.)

Phone: 570-323-5001 Fax: 570-323-0009

Metrics Inc.

Phone: 252-752-3800 Fax: 252-758-8522

E-mail: [emailprotected] Web Site http: 1l~w.metricsinc.com

PE Informatics 3822 N First St San Jose, CA 95134

Phone: 408-577-2200 Fax: 408-894-9307

Web Site http: ll ~ ~ . p e i n f o r m a t i c s . c o ~

Perkin-Elmer Corp

761 Main Ave Wilton, CT 06859

Phone: 203-762-4000 Fax: 203-762-4228

E-mail: info~perkin-elmer.com Web Site: htttx llwww.~erkin-e~er.com

Pham-Eco Laboratories

128 Spring St Lexington, MA 02421

Phone: 781-861-9303 Fax: 781-861-9386

E-mail: boast@pharme~o.com Web Site: httD:ll ~ w . ~ h a r m e c o . c o m

ScienTek Software, Inc.

P.O. Box 323 Tustin, CA 92681-0323

Phone: 714-832-7435 Fax: 714-832-7435

Thermo Labsystems

100 C u ~ i n g Center s Suite 407J Beverly, MA 0191 5

Phone: 978-524-1400 Fax: 978-524- 1244

P.O. Box 4035 Greenville, NC 27836

trieval of sta~ilitydata ting of data in multiple formats dation of data raphical presentation of data aintenance of schedules eneration of test assays scheduled

E-mail: sales~labsystems.~om Web Site httu: l1~w.labsvstems.com

Statistical evaluation of the data Linear and multiple regression Expiry data evaluation Shelf life projections Analysis for pooling of batches ultivariable search capability ojections for current and future workload aintenance of active batches ance of inactive batches Classification of studies by type mples/data (e.g. bar codes) Provision for multiple levels of authorizations aintenance of audit trails Amendments for data Validation capabilities Inventory management Back-up security provision The following parameters should be considered as anessential listof attributes for initiation of stability studies inthe software system. It should benoted that many of these itemsare part of the information for submission to the regulatory agencies, while others are for management of the stability function within the organization. rug substance or drug product name ate stability initiated urpose of stability study afety handling precautions tability study number or code ormulation /product number or code atch number or lot number Storage conditions ocation of samples ests to be performed nternal test specifications or target egistration test specification nalytical method to be used Laboratory performing the test uration of the study ternal accounting charges Osage form and strength Labeled strength Theoretical strength (including overage) atch type and size

ate of manufacture te of manufacture rug substance batch number rug substance manufacturer g substance manufacturing site tainer description or composition Closure description or composition iption or composition

hould all organizations considerdeveloping or acquiring the computer system for stability information? There is no simple answer.It depends on the needs of the or~anization,the volume of stability data generated, the computational demands, and the need for dissemination of information. Development or acquisition, establishment and implementation of computerized stability systems are fairly expensive propositions. Not every organization can justify the cost associated with the development, training, validation, maintenance, upgrades, and operational needs. In reality, stability data is not much different from other data, such as inventory, payroll, or employee information data. Itdoes, however, require careful consideration of compliance with goodlaboratory practices and/or good manufacturing practices regul ons, especially if the systems are used as aprimary database for making decisions ne can apply general principles involved inthe development of information systems to stability information systems. The terms data and formation are often used interchangeably, but they refer to two distinct concepts. ncewe are dealing with stability data and information, let usexamine the concepts from stabil representations of events or concepts. ents, and in-vitro dissolution resul of data. Information on the other hand is subjectiveand relative. While data and information are separate concepts they are distinctly related; info~mationis produced from data. The computation of an expiry date, either from the graphical representation of stability data or from a rigorous statistical analysis of the data, is an example of information, n simple terms, data are the raw material from which the finished product, information, is produced. ~ l t h o u g hdata are the key ingredient used to produce information, not all data provide relevantand timely information. The sheer volume of the data can be a burden to an organization. Collection the data justfor the sake of collecting is a most nonproductive useof resources. ata collection must add value to the or~anization. he only way to reduce this burden of unnecessary information is through implementing an effective and efficient user-oriented information system. The purpose of information is to increase the knowledge levelof the recipient. S, the information should provide a clear picture of what has mation serves another very important function for the person who receivesit. Thatfunction is to reduce the variety of choices and the uncertainty related to thesechoices. nformation is critical in the decision-makingprocess.

A single stability result for a 6 months time point for a given batch may not be meaningful, but when it is viewed in conjunction with initial results for the same batch, it becomes useful information whether any significantloss from initial result has occurred or not. er

ow are the data converted into information? Several specific data operations are Involved in the con~ersionprocess. Any one or combination of these operations can produce information from data. The data operations are: Capturing: This operation refers to recording of data from an event. example, a reading from a chart pape a pH meter, or a li chro~atographare examples of capturing Verifying: This pertains to validating data to insure that it was captured and recorded correctly. This is critical operational step. This assures the integrity of the data. Example includes the review by another person to verify the accuracy of the data. In computer data entry, the verification often means rekeying in the data in order to match the earlier entry. Classifying: This operation places data elements into specific categories, which provide meaning for the user. Examples of this could be classification of stability data for solid dosage forms, sterile products, and topical dosage forms. One may classifythe solids data by break in^ up further into tablets, soft gelatin capsules, hard filled capsules, etc. Arranging (sorting): This operation places data elements in a specified or predetermined sequence. For example, arranging the stability data for a given lot by age of the samples, or by temperatures. Summarizin~:This operation combines or aggregates data elements ineither a mathe~aticalor a logical sense. For exam le, the stability data could be summarized for a product type. ~ i t h i anproduct type, one could summarize by different formulations. Another wayis to summarize the data by divisional responsibility, e.g., production versus research. S the name implies, it ~ ~ r f o the r ~arith s for example calculation of percent of 1 raw data, to determine if the assay results are belo specified limit, or to determine if it meets a test or not. toring: This operation simply places data into some storage media such as a file drawer, a computer disk, microfilm, or magnetic ta can be retrieved. etrieving: This operation means searching out and gaining accessto specific data elements from the storage medium. eproducing: This operation duplicates data from one medium to another. A common example would be a back-up copy of a magnetic tape or disk for further processing or security purposes or for a disaster plan. C o ~ u n i c a t i n g This : involves transferring data from one place to another. The examples could be the transfer of data onto a computer monitor or to a printer. The ultimate aim of all informat~onsystems isto c o ~ m u ~ i c a t e or disseminate information to the final user.

ns These data operations are involved in any management of data. They are independent of the data processing method, whether it is done manually or through the use ofa computer. The decision to use one of these methods depends on the economic considerations, the processing requirements, and the performance factors related to the data processing method. Data processing requirements are based on Volume of the data: Volume simply meansthe number of data units that must be processed to achieve an information goal. For a stability information system, each data unit directly comes from each stability result. The total volume in any organization will depend on the number of stability batches, and the average number of tests performed per batch in a given period. Complexity: Complexity refers to the number of intricate and interrelated data operations that must be performed to achieve an information goal. In the case of a stability system, oncethe results are entered, they must be checked for any systematic types of errors, added to the proper batch, inform the user, keep track of results in the file, identify anyresults that have not been entered, and so on. Processing timeconstraints: Time constraints can be defined as the amount of time allowed or acceptable between when the data are available and when the information is required. In today’s information age, all of us want more and more information and we want it sooner rather than later. Computational demands: These are a unique combination of both volume and complexity. For a stability system, the computational requirements could be the calculation of shelf-life based on a zero order or a first order, or the application of the Arrhenius relation to multitemperature stability data.

So far we have identified the processing requirements for a stability information evelopment of the system requirescareful consideration of many performance factors as follows. Initial investment: This is just the initial expense of acquiring any materials or hardware and software required for processing. Setup: This is the expense required to capture initial data for subse~uent processing. Conversion: The one-time expenseof processing data with the new computer method. Skilled personnelrequirements: This simply meansthe education and training of the individuals involved in operating and using the system, Variable cost: The cost of a data unit based on the total volume. In general, computers allow greater efficiencyin this area. The greater the volume of data that needs to be handled, the cheaper it becomes per unit cost. odularity: This represents the ability to increase or decrease processingcapability to match the requirements for processing. For example, if we need one data entry computer terminal to process 5,000records a day, two computer terminals will provide for data entry up to 10,000 records per day.

Flexibility: The ability to change the processing procedure to satisfy new or changing requirements. No system remains static for long. ell-designed systems provide excellent flexibilityto modify the system to meet changing needs of the organization. Versatility: The ability to perform many differenttasks. This is wherethe computer excels as it can handle multiple tasks. rocessingspeed:This represents the time required to convert inputs to outputs. This is not a problem as the processing speedsof today’s computers continue to improve in an exponential manner. Computational power: The ability to perform complex mathematical operations. This is very inefficient in the manual system. This is excellent for the computer. Automatic error detection: The ability to identifyprocessing errors. The human mind is capable of detecting errors, but we have a fatigue problem. On the other hand, the computer never gets tired. This capability in the computer systemprovidesexcellentdividends in terms of continuous monitoring. ecision-making power: The ability to chose among various alternatives in order to continue processing. Againthe speed with which the manual system would workin this task is slow. The computer has high capability of making decisions along the way if the system is well designed. System degradation: The level to which the processing system is degraded because of breakdown or unavailability of a component. puter breaks down, the system is completely down. The design and developmental consideration of a computer system for stability information mana~ementinvolves the following major steps. SYSTEMSANALYSIS. Here we need to focus on current capabilities of present systems and determine what the new systems should do. In this phase, we have to look at users and any problems faced by the users. hat do they need? Are there any problems with the system as it is being used? So imes the users feel that they need the information because the predecessors have done so. Often, the users do not know what can be provided. Systems scope covers what capability the system is supposed provide. Should it include only the scheduling of stability? Should it provide computations? Finally the requirements for the system are formulated through the system design phase. GENERAL SYSTEMS DESIGN. In systems analysisthe focus was withwhat the ystem isdoing, and what the new system should bedoing to meet the requirements. n the system design phase one has to concentrate on how the system is developedto meet theserequirements. In the broad design phase questions like t if? Why not? need to be addressed. A rough sketch or an outline of S developed. The input and output of the system needto be closelylooked ds to determine whether there is more than one way to get what is expected from the information system. SYSTEMS EVALUATION AND JUSTIFICATION. In this phase one has to evaluate the impact of the propos system on the employees. The system should be such that it assists the user, rat than requiring the user to changesignificantly the

way he or she operates. should not expect the user to be a genius in order the system. The system Id serve the needs of the use the evaluation of cost-effectiveness shouldbe undertaken. the indirect costs? has to determine the fixed costs, which independent of con ng needs as well as operational CO plexities and the volume of data. Again, one has to CO just as in the system design. DETAILED SYSTEMS DESIGN. This isthe most critical ste in the design phas representing the fine-tuning of the broad design. This the is stage t which the detaile specificationsof the systems are developed. Onehas to look at not only what the input would be but also how the input would be prepared, what t es of forms are needed for input, and what different types of output are needed. at type of data entry method is most suitable, and who would enter the dat p asking the users for more and more infor ring this phase every aspect of the system is cl Often people overlook the importance of a well-de output of the data. The input forms should be designed Complicated forms only cause confusion. The form sho serves the needs of the person who l record the data assist the person in data processing. processing people, completely ignor Today’s computer printers have come a long way from the fixed font ty printer of earlier years. They allow printing of information in a nice layout exceptional clarity if enough thought process has gone into designin and output forms. hysical layout as well as order of items on the forms shoul follow reasonable logic that serves the user’s needs, SYSTEMVALIDATIONANDIMPLEMENTATION. his is the final phase of sysuring this phase, one should familiarize the users about the system by proper training and education. The more the usersknow the the system, the more they will be able to utilize all the capabilities. the users are the ones who will continually come up with a list of improveme efore finalimplementation, extensive validation of the system ne n. Next comes the testing of all of the individual components or programs of the system. All of the inputs should be used to pro output. One has to check the output for accuracy and reliability. Vali puterized systems requiresa lotof resources. Ifnot properly planned, validation can consume more resources than that required for the development or imple udihandojo, 1997, 1998; amw ways et al. 1996; ~choenau Finally the system requires continuous maintenance. without proper maintenance. aintenance does not only mean to keep the syste going but also to keep the system up to dateto serve the changing needsof the user. There is nothing more frustrating than using a system designed using outmoded standards.

Allen, I?. V., Dukes, G. R., Gerger, M. E.Determination methodology. Pharm. Res. 8(9):1210-1213,1991.

of release limits: general

udihandojo, Rory. Computerized system validation: a concept approach in the preparation of a validation plan document. Phamaceutical Technology 21 (2):70, 1997. Budihandojo, Rory. Computerized system validation: preparation of a design speci~cations document. Pharmaceutical Technology 22(3): 150, 1998. rew, H. D., Thornton, L. K., Juhl, W. E., Brower, J. F. An FDA/PhRMA interlaboratory study of the International Conference on Ha~onization'sproposed photostability testing procedures and guidelines. Pharmacopeial Forum 24(3):6317, 1998a. Drew, H. D., Brower, J. F., Juhl, W. E,, Thornton,L. K. Quinine photochemistry: a proposed chemical actinometer system to monitor UV-A exposure in ~hotostabilitystudies of pharmaceutical drug substances and drug products. Pharmacopeial Forum 24(3):6334, 19988. Fairweather, W. R., Lin, T.-Y. D., Kelly, R. Regulatory design and analysis aspects of complex stability studies. J. Pharm. Sci.84:1322-1326,1995. Food and Drug ~dministration(FDA). FDA guide to inspections of pharmaceutical quality control laboratories. July 1993. FDA. Content and format of investigational new drug applications (INDs) for phase 1 studies of drugs, including well-characterized, therapeutic, biotechnolo guidance for industry. Center for Drug Evaluations Research Biologics Evaluation and Research (CBER), November 1995. FDA. SUPAC-IR: immediate release solid dosage foms: scale-up and post-approval changes, chemistry, manufacturing, and controls, in vitro dissolution testing, in vivo bio~~uivalence documentation. CDER, November 1995b. FDA. S ~ ~ A C - I R / M R immediate : and modified release solid dosageforms, manufacturing equipment addendum, draft guidance. April 1998a. FDA. Guidance for industry: stability testing of drug substances and drug products, draft guidance. June 1998b. Grimm, W. Drugs Made in Germany 28:196--202,1985. rugs Made in Germany 29:39-47, 1986. Haynes, J. D. The world with virtual temperatures for product stability testing. J. Pharm. Sci. 60:927-929,1971. I~ternationalConference on Harmonization (ICH). Q1A stability testing for new drug substances and products, September 1994. .Q123 photostability testing of new drug substances and products. November 1996. Nordbrock, E. Statistical comparison of stability study designs. J. Biopharm. Stat. 2(1):91-113,1992. Sager, N., Barn, R. G., Wolters, R. J., Layloff, T. Photostability testing of pharmaceutical products. Pharmacopeial Forum 24(3):633 1, 1998. Samways, K., Sattler, L., Wherry, R. Computerized system validation: one company's approach to assigning tasks and responsibilities. Pharmaceutical Technology 20(8):45, 1996. Schoenauer, Ciaran M., Wherry, Robert J., et al. Computer systemvalidation-staying current: security in computerized systems PMA's computer systems validation comPharmaceutical Technology 17(5):48, 1993. .H., ed. Photostability of Drugs and Drug Formulations, London: Taylor and Francis, 1996, Yoshioka, S., et al. ~ u i n i n eactinometry as a method for calibrating ~travioletradiation intensity in ~ight-stabilitytesting of pharmaceuticals. Drug Development and Industrial Pharmacy 20(13):2049-2062,1994.

Biogen, Cam~ridg~, ~assachusetts Cubist Pharmaceuticals, Inc., C a m b r i ~ g ~ , ~ a s s a c h u s ~ t t s 553 554 554 554 558 560 560 570 573

5. Conclusions

eferences

573 573

Several differences exist betweenconventional small molecular weight compounds and polypeptides, many of which impact their predicted stability profile. peptides consist of a regularly repeating backbone with distinctive side chainsthat interact with each other to contribute to the three-dimensional structure of the protein (see Fig. 1). Typically, small molecular weight compounds are either linear or cyclical in nature, and their size prohibits extensive intramolecular bonding. Thus the majority of co~ventionalsmall molecular weight compounds do not exhibit higher level structures found in polypeptide molecules. It is possibleto make replacements of similar amino acids withina polypeptide with little impact to the biological activity and structural stability of the molecule. 3

Schematic representation of a pentapeptide.

Therefore chemical changescan be made to the compound without loss of activity. This is in distinct contradiction to small molecular weight compounds. Loss of chemical integrity or conformational structure of a polypeptide can affect the molecule’s activityat several levels.As with conventional small molecular weight compounds, potency can be decreased or entirely lost due to degradation of the molecule. In addition, protein degradation can result in acquired immunogenicity, altered pharmacokinetics, and protein self-association. This chapter will discussvarious methods that cause protein degradation and will provide several methods for detecting these changes.

olype~tidesare generated by the condensation of two amino acid residues.As illustrated in Fig. 2, the peptide bond is formed between the a-carboxyl group of one amino acid and the a-amino group of another. It is noted that hydrolysis of the peptide bond is favored over its synthesis. Therefore the formation of peptide bonds requires input energy.

Provided in Table 1 is a list of the 20 naturally occurring amino acids that make up mostpolypeptides and proteins. Theseresidues can beclassified further based on their side chain functionality. With the exception of glycine, allamino acids contain a chiral carbon and are thus restricted in their conformation.

Formation of peptide bond.

ilit

Twenty Naturally Occuring Amino Acids with Corresponding Abbreviation and Mode(s) of Degradation Abbreviation Residue 3-Letter

l-Letter

Alanine Arginine Asparagine

Ala Arg Asn

A R N

Aspartic acid

Cysteine

Glutamic acid Glutamine

Glu Gln

Q G

Glycine Histidine Isoleucine Leucine Lysine Methionine Phenylalanine Proline

His Ile Leu LYS Met Phe Pro

H I L K M F P

Serine Threonine Tryptophan Tyrosine Valine

Ser Thr TrP TYr Val

S T W IC V

2.2.1.

Potential mode(s)

of degradation

Relatively stable Relatively stable Deamidation, racemization isomerization Hydrolysis, racemization, isomerization Oxidation, p-elimination, racemization, disulfide exchange Relatively stable Deamidation, racemization, isomerization N-terminal location promotes diketopiperazine formation Oxidation Relatively stable Relatively stable Relatively stable Oxidation Relatively stable N-terminal location promotes diketopiperazine formation p-elimination, racemization p-elimination, racemization Oxidation Oxidation Relatively stable

~ y d r o ~ ~ oAmino b i c Acids

Hydrophobic amino acids contain either aliphatic or aromatic side chains (see Figs.3 ost hydrophobic residues are located in the interior of the polypeptide, thereby making them relatively unreactive. Although this group of amino acids is primarily hydrophobic in nature, the nitrogen in the indole ring of tryptophan and the phenolic group in tyrosine can H-bond with other residues. In addition, the proximity of the methyl groups in valine and isoleucine to the main chain restricts polypeptide conformation in regions where these residues are present. 2.2.2. Neutral- olar AminoAcids

Neutral-polar amino acids contain polar groups that are not readily ionizable (see Fig. 5). Asparaginyl and glutaminyl residueshave the propensity to deamidate to aspartic and glutamic acid, respectively. In addition, when centrally located, asparagine and glutamine often H-bond with amides. Hydroxyl groups present in the side chain of serine and threonine allow themto hydrogen bond with the main

CH3

L~u~in

Hydrophobic-aliphatic amino acids.

yrosin Hydrophobic-aromatic amino acids. chain thus providing the opportunity for elimination and racemi~ation. ine residues are primarily located within the interior region of polype are highly susceptibleto oxidation. Tryptophan and t sine also have the potential to oxidize, though to a lesser extent than methionine e sulfhydryl groups within each of the cysteine residues react together to form disulfide bridges. 2.2.3.

AcidicAminoAcids

Acidic amino acids contain side chains that arenegatively charged at p~ysiological (see Fig. 6). Acidic amino acids are typically located at the polypeptide surface and are frequently the active site for enzymatic reactions. Aspartyl residues are often susceptible to hydrolysis, racemization, and isomerization while glutamyl residues are relatively stable. 2.2.4. BasicAminoAcids

asic amino acids either possess a neutral (histidine) or positive (lysine, arginine) charge at physiological pH (see Fig. 7). Histidine is often involved in enzymatic reactions and can be prone to oxidation. When located in the interior portion of a polypeptide, basic amino acids (lysine, arginine) can be involved in electrostatic interactions with acidic residues (i.e., aspartic acid, glutamic acid).

H2N-CH-C-OH 0

H2N-CH-C-Ot-1 H2N-CH-C-OH H2N-CH-C-OH

I I CH3

CH-OH

p"'

Serine

Threonine

YH2

p"'

CH3

YH2

Methionine

Cysteine

0 0

H2N-CH-C-OH

PH2

c=o

NH2

~ H 2 ~ H 2

c=o

NH2

Glutamine

Neutral-polar amino acids.

Acidic amino acids.

2.2.5. Conform~tion Influencing Amino Acids

Glycine and proline (see Fig.8) are two residuesthat play a major role ina protein's conformation. Thesetworesidues,however,behave quite differently from one ue to its achiral nature, glycine isable to adopt conformations forbidden by other residues. In addition, glycine is the smallest of the 20 amino acid residues. Thus, glycine is highly flexibleand is often located at turns in the protein. Proline, on the other hand, is very rigid because its side chain nitrogen is covalently linked

ott

CH*

Lysine

Basic amino acids.

1 l

H-c-OH H

lycine

rolin

Conformation influencing amino acids.

to the main chain of the polypeptide. Therefore regions containing proline residues are restricted in their conformation. Interestingly, the presence of either glycine or proline in an N-terminal position promotes diketopiperazine formation.

As noted in the introduction, the primary factor that differentiates polypeptides and proteins from conventional small molecular weightcompounds is the varying levels of structure that polypeptides and proteins exhibit. Changes at any level ofthe protein structure can impact the molecule’s activity. Protein chains fold spontaneously into their native (i.e.,biologicallyactive) conformation, most often with the hydrophobic residues interior to the protein and the charged residues on the surface. rotein denaturation occurs when a polypeptide loses its higher level structure and often results in loss of biological activity. 2.3.1.

Primary

The primary structure (Fig. 9) of a polypeptide is composed of its amino acid sequence and the location of any disulfide bridges. Thus all of the covalent connections within a polypeptide contribute to its primary structure.

Protein structure: (a) primary; (b)secondary; (c) tertiary; (d) quaternary.

2.3.2.

Secondary

Secondary structure refers to the localized foldingor shape of a polypeptide. Amino acid sequence (i.e., primary structure), residue alignment, hydrogen bonding, and disulfide bridges all contribute to the secondary structure of a polypeptide. Therefore secondary structure results from the interaction between adjacent residues, or those within close proximity of each other. Examples of periodic secondary structure include the a-helix (as depicted in Fig. S), the P-pleated sheet, and the P-turn. 2.3.3.

Tertiary

A polypeptide’s tertiary structure imparts three-dimensionalityon the molecule. It is derived from the spatial arrangement of nonadjacent residues. In other words, tertiary structure is the result of the interactions between the residues’ side chains and their resultant orientation. 2.3.4.

Quaternary

Proteins containing more than one polypeptidechain, or subunit, exhibit quaternary structure. The quaternary structure of a protein is a result of noncovalent binding of its subunits. The manner in which the subunits are aligned witheach other contribute to its quaternary structure.

hough polypeptide and protein stability have been se arated into chemical and ysical modes, there is obviouslya significant relationsh between conformational sta~ilityand the chemical integrity of each molecule. This relationship is of par .. lar importance to the understanding of the mechanisms of protein inactivation. turbation of secondary or tertiary structure can lead to exposure of previously bune amino acids, facilitating their chemical reactivity; alternatively, chemical changes can lead to loss of co~ormationalstability. ~ u m e r o u sexcipients incorporated into pharmaceutical products can also ffect protein stabili example,reducing sugars can re roups to result in th tion of a SchifPsbase and subs n addition, several surfactants, such as Tween 20, Tween mers, can cause the oxidation of susceptible amino acids present in these materials that often increase on storag of some of these surfactants are now being made available with minimal peroxide levels, and some are also being produced with antioxidants, such as butylated hydroxytoluene (BHT). Antioxidants themselves should not be incorporated into f o ~ u l a t i o n sof proteins that contain disulfide bonds, sincetheymay contain reducing agents that will destroy this bond, As many of the chemical reactions involving polypeptidesand proteins are catalyzed or facilitated by metal ions, attention should also be made to levels of these in formulated products. *

ithin the context of this chapter, chemical instability refers to the formation or destruction of covalent bonds within a polypeptide or protein molecule.These changes alter the primary structure of the protein and can also impact higher levels of its structure. ~ o m m o ncauses of chemical instability of polypeptide molecules include deamidation, oxidation, and cystine destruction/disulfide exchange. onenzymatic deamidation of asparagine residues occursin many proteins and peptides and is a major source of spontaneous degradation and loss of amino acid sequence homogeneity. While glutamine residues can also undergo deamidation, they are muchlesslabile to this reaction than asparagine. Asparagine residues deamidate to aspartate or isoaspartate, changing positive charge to negative. Althou~hthe reaction is generally associated with conditions of neutral to basic ,mechanistic studies have shown that deamidation can also occur in acidicpH. The neutral-to-basic pH deamidation reaction is intramolecular and occurs via a five-membered ring succinimide intermediate (see Fig. 10) by attack of the side chain amide on the backbone of the peptide carbonyl. The s ~ ~ ~ i n i mring i d eis labile to base and readily hydrolyzes in two possible ways, yielding either aspartic acid in a normal a-linked form and/or isoaspartyl characterized by an atypical bond between the P-carbonyl of aspartate and the “nitrogen of the neighboring amino acid. Although, in general, proteins are less susceptibleto deamidation under acidic conditions, deamidation by direction hydrolysis, rather than via imide formation, can occur as pH continues to decrease.

~eamidationof L-asparagyi residue by succinimide formation.

~usceptibilityof asparagine or glutamine to deamidation is highly se~uence and conformation dependent. Studies with model peptides show that asparagine, proceeded by glycine and serine residues,is particularly labile due to the decreased rance of succinimide formation by the C-terminal residue (1). In contrast, bulkyresiduessuch as proline reduce the reactivity of asparaginyl residues. T~erefore, alt~ough conformationally rigidregions of a protein moleculemay inhibit deamidation at labile asparagine residues, the presence of a sensitive aspara-

Several Polypeptides and Proteins that are Subject to Deamidation Protein/peptide Reference residues

Affected

Insulin

Asn~3Gln

Brange et al. 1992 (2); Berson and Yalow 1996 (3)

Insulin

Asn~21(C-terminus)

Strickley and Anderson 1996 (4)

Adrenocorticotropi~ Asn~tjGly hormone (ACTH)

Bhatt et al. 1990 (5)

Human Asnl49Asp growth hormone (hGH)

Becker et al. 1987 (6); Johnson et al. 1989

Inter1eukin-la (IL-1a) Asn36Asp

Wingfield et al. 1987 (8)

Human growth hormone releasing factor (hGHRF)

Asng Ser

Friedman et al. 1991 (9)

CD4

Asn52Asp

Teshirna et al. 1991 (10)

Tissue plasminogen Activator (tPA)

Asn37Ser9Asn~g Gly, Asnl77Ser

Paranandi et al. 1994 (1 1)

Hirudin

Asn33Gly, Asn53Gly

Tuong et al. 1992 (12); (13)

(7)

gine sequencein a particularly flexible region (e.g., flanked by glycine or serine) may enhance the susceptibility of this site to deamidation. A s noted in Table 2, deamidation has been documented in a number of proteins and can impact significantly the molecule's activity, Deamidation can make proteins more susceptible to proteases and denaturation. Increased susceptibility to proteases induced by deamidation can reduce the in vivohalf-life of the molecule.An additional problem posed by deamidation of asparaginyl residues is the formation of isoaspartate. Isoaspartate residues can dramatically affect the activity and conformation of a protein (14.) and may also increase the immunogenicity of certain proteins (15). Note that isoaspartate can also arise from direct iso~erizationof aspartate. This reaction, which is generally more difficult to detect as no change in charge occurs, is optimal at p eamidation in pharmaceutical products can be detected by charge or molecular weightchanges by direct monitoring of the formation of succinimide or isoaspartic residues. thods used to identify deamidation include peptide maps, capillary electrophoresis,isoelectric focusing,and enzyme catalyzedradiolabeling of the isoaspartyl sites (n.b., commercial kits are available). Formulation approaches to minimize deamidation generally involvemaintaining conditions oflow pH where deamidation isless desialylation of glycosylated proteins is prevalent at low with proteins that are subject to both deamidation and desialylation, optimization studies to balance these conflicting objectives should be performed. As deamidation may also be affected by buffer composition, studies evaluating polypeptide stability in varying buffers may also provide valuable information.

3.1.2.

~xid~tion

ethionine, cysteine, tryptophan?and tyrosine residues are all susceptible to oxidation. Cysteine and methionine oxidation, however, are more common and have hionine oxidation, induced by exposure to air,intense fluorescent light, or ts in conversion of the thioether to sulfoxide as shown onditions result in the formation ofmlfone. ring fermentation or cell culture and purifi great care should be taken when modifying such susceptiblemethionineresidues are present. noteworthy, however, that several new approaches to remove oxidized material during purification are being developed (19). 3

oxidation under mild (dilute H202)and strong (performic acid) conditions.

0 ethionine oxidation is also of particular concern during the storage of both liquid and lyophilized formulated products (20-23). The susceptibility of methionine to o~idation has been shownto be dependent on the relative exposure of the residue (24,25). ~lterationof pH, ionic strength, and solvent polarity can change both the rate and the extent of methionine oxidation. ~ l t h o u g hthe impact of oxidative modificationson proteins is variable, studies with recombinant human leptin show that the four methionine residues in the molecule are notequally susceptibleto oxi~ation and that the resulting oxidativeproducts each have a different magnitude of effect on maintaining leptin’s structural and biological activity (26). Selective oxidation of another protein, recombinant human granulo~ytecolony sti~ulatingfactor (rh CSF), by hy~rogen pero~ide generated four different modified forms of the mole one to four modified met hi on in^ residues. A11 oxidized forms of G-CSF retained a c o n f o ~ a t i o n astrucl ture similar to thatof the native molecule, but only oneretained all of its biological function. In addition, site-directed mutagenesis with leucine replacement of methionine produced increased stability and retained the biological activity of the molecule (24). Thus chemical modification to G-CSF was made without impact to its activity. Antithrombin provides another interesting example of the impact of methionine oxidation on biological act contains four methionine residues; two adjacent residues near the reactive site loop cleaved by thrombin, and exposed tworesi et20) that border on the heparin-binding region. In forced oxidation studies with h ~ d r o g ~peroxide, n the methionine residuesat positions 3 14 and 3 15 werefound to be highly susceptible to oxidation, but their oxidation did not affect the molecule’s thrombin-inhibitory activity or heparin binding. Significantoxidation of methionine residues at positions 17 and 20 only occurred at higher peroxide concentrations, at which point heparin affinity was decreased.Structural studies indicate that highly oxidizedantithro~bin less able to undergo the complete conformational change inducedby heparin (27). dditional studies evaluating the oxidation of methionine residues are provided in *

Several Polypeptides and Proteins that AreSubject to Protein/peptide ~ntistasin Cystatin C

ethionyl Oxidation

Reference Przysiecki et al. 1992 (18) erti et al. 1997 (19)

R e ~ o m b i n a human ~t monoclonal antibody HER2 (rhu~AbHER2)

Lam et al. 1997 (21)

~ranulocyte c~lony-stimulating factor (C-CSF)

Lu et al. 1999 (25)

Leptin

Liu et al. 1998 (26)

Antithrombin

Van Patten et al. 1999 (27)

Epidermal growth factor (ECF)

C e o r g ~ - ~ a s ~ i m e et n t al. o 1988 (28)

Adrenocorticotropin hormone ( A ~ ~ H )

Antonini and Spoto 1988 (29)

ptide maps are convenient approaches for detecting methionine oxidation, peptide containing the sulfoxide being identified by mass spectroscopy LC) is often successful at separating oxidized forms and is also used to identifyoxidized products by coelution of hydrogen peroxide stressed molecules. Approaches to minimize the formation of oxidative products are based on the ants to the formulation and the manipulation of environmental or temperature adjustment). Antioxidants such as methionine, sodium thiosulfate, catalase, or platinum can prevent methionine oxidation, presumably by acting as free radicals or oxygen scavengers (21). Oxidation of cysteine residues ispromoted at both neutral and elevated pH and can result in aggregation or disulfide scrambling (see Sec. 3 1.3). Thus maintaining low p the incidence of methionine oxidation. A recent study reports the identification of tryptophan oxidation products (hydroxytryptophan, N-form e, and kynurenine) in intact bovine alpha-crystallin protein (30). idation of tryptophan residues has also been reported to occur in luteinizormone-releasing factor (LRF), somatostatin, valine-gramicidin A, and AC 3.1.3. Cystine ~ e ~ t r u c t i o and n Thiol-~isulfi~e

Cysteine residuescontain thiol groups that can be oxidized to form disulfide bonds (cystine residues). These bonds are naturally occurring cross-links that covalently connect protein or polypeptide chains either intra- or intermolecularly. can be formed either by direct oxidation of cysteine or by thiol~isulfideinterchange and are often catalyzed by trace copper and iron, the relative stability of a reduced eine being dependent on the redox potential of the protein’s environment. Each unit change towards the ic has been reported to decrease the oxidative tyof thiols by approxi ly an order of magnitude (33). isulfide bonds general1stabilizingfolded forms of proteins, but can be to destruction under certain conditions such as reducing potential, heat, and ~struction of disulfide bonds can occur at elevate e Fig. 12). For example, at pH 9 and 50°C’ formation cross-linksin hirudin resulted in the molecule’slossof activity (34) scrambling has also been documented in human insulin-like growth fa (23) whereformation of intermolecular disulfides resulted in dimerizatio protein in which atypical disulfide bonds are formed is acidicfibroblast growth factor aFGF normally contains three reduced cysteine residues and no disulfide owever, copper is able to catalyze the formation of an intramolecular disreby causing inactivation of the aFGF molecule (35). The presence of atypical disulfide bonding can often be suggested by running r nonreduced sodium dodecyl sulfate polyacrylamide gel electrophoresis ( ); and Ellman’s reagent is widely used for the determination of free thiols (36). Methods used to study reactions between free thiolsand disulfide bonds in more detail include various proteolytic digestions,peptide mapping, partial reduction, and assignment disulfides of by N-terminal sequencing and matrix-assisted laser desorption ionization ( ALDI) MS. The use of the majority of these methods is documented in a recent study on the determination of the disulfide structure of tumor necrosis factor binding protein (37).

L-am~noacid residue

C a i~natnei~oend i a t e

elimination) D-amino acid residue

or ar~aticresidue

1 aliphati~

~-Eli~ination and racemization in alkaline media.

nimize the destruction ofdisulfides and formulations and avoiding mber of antioxidant excipients have signifiof therapeutic proteins obviously reduces eactions, but it ispossible to facilitate odification due to dehydration of the molapproach to resolving potentia~instability hiols in the presence of disulfides is the site-specific substitution residues. Thishas been done with a human interbeta-lb, which contains one disulfide bond and

other less common or less documented bond cleavage reactions also ides and proteins and are discussed below. olysisat A s ~ ~ r t Acid ic of hydro1 ofsis the

idues ~ n d e Acidic r ~on~itions -terminal peptide bond of S hydrolysis, documented primarily at te at least 100 times greater than any

spartyl peptide

Hydrolysis of aspartyl peptide in dilute acid.

'other amino acid (40), and the Asp- ro bond is known to be particularly labile, possibly due to either an enhanced a l p isomerization of the aspartic acid residue or the more basic nature of the proline nitrogen (41). Cleavage of an Asp-Ser bond has been reported on storage of recombinan (rhEGF) under the accelerated conditions of

L-aspar~ginyl residue

C-terminal succinimide frag~en~

Newly exposed N-terminus

C-Terminal succinimide formation of asparginyl residue.

t i

~ i ~ e t o ~ i ~ e r aformation. zine

3.1.4.2. ~ - ~ e r ~ i n a l ~ u c ~Formation i n i ~ i d at e as para gin^ ~ e s i d ~ e ~ Succinimide formation as occurs during deamidation reactions can also result in the spontaneous cleavage of the peptide bond at asparagine residues. In this case, the side chain amide nitrogen attacks the peptide bond to form a C-terminal succinimide residue and a newly formed amino acid terminus (43). See Fig. 14. 3.1.4.3, ~ i ~ e t o p ~ e r a z i n eF o r ~ a t i o n arrangement of the N-terminal dipeptide can result in the formation and cleavage a cyclic diketopiperazine at high pH (see Fig. 15). Proline and glycine at the ~ - t e r ~ i n u(44) s promote this base-catalyzed reaction. The kinetics of diketopipera~ineformation has been studied using model peptides. In one study, degra~ationdue to the formation of diketopipera~inefollowed pseudo-first order kinetics and showed a significant dependence on p 3.1.4.4. ~ n ~ y ~ aProteolysis tic andAutolysis Contaminating proteases present during fermentation, cell culture, or purification can result in the cleavage of recombinant proteins. Optimi~ationof processing conditions and addition of carefully selected protease inhibitors can minimize this occurrence, Storage of purified proteases may also lead to peptide bond cleavage; tissueplasminogen activator (tPA), a serine protease, contains an activation e site at Arg275. Even after autolysis, the two chains of the clip held together by a disulfide bond between residues 264and 395. of the clipping by site-directed mutagenesis of the labile protein was attempted by replacing the Arg275 residue with other amino acids. The resulting one-chain

ilit

tein

A, however, exhibited diminished cofactor (46).

activity in the absence of a fibrin

3.1.4.5. AlternativeCovalent Agg~egation Although the most commonform of covalent aggregation is due to the f o ~ a t i o of n disulfide bonds, aggregation can also occur via free amino, carbonyl, carbox hydroxyl functional groups with ester or amidelinkages, for example. aggregates have been dQcumented for insulin (47,48)and were also found during the analysis of relaxin (49). nd ~esialylation

There are a number of glycosylated therapeutic proteins that have sugar and sialic acid molecules covalently attached to the peptide backbone. Runkel et al. report beta produced in mammalian cell culture as a glycosylated protein, identhuman molecule, has greater stability to aggregation than the corresponding protein produced by bacterial fermentation in the nonglycosylated form (50). enerally, bonds attaching sugars to proteins are stable, and any modifications carbohydrate structure are due to process changes during the manufacture of esialylation can however occurat acidic p to carbohydrate structure, or changes in the glycosylatio can result in significant changes inthe pha~acokinetics sialic acid content has been shown to be responsible fo of highly purified human pituitary luteinizing hormone isoforms (51). Changes in glycosylation can be detected by various gel methods including fluorophore-assisted carbohydrate electrophoresis (FACE) and sialic acidcontent can be detected by measurement of free sialicacid, New methods to quantitate oligosaccharides are being developed. For example, high res of normal phase high-pH ion-exchange chromatography and normal phas methods combinedwith can provideeffectivetechnology for analyzing a wide repertoire of oligosaccharides structures (52). radation of Proteins

th ionizing and nonionizing radiation can cause protein inactivation. osine, and cysteine are particularly susceptible to nonionizing radiation, such as ultraviolet (UV) light. The absorption of photons leads to photoionization and the formation of photodegradation products through either direct interaction with amino acid or indirectly via various sensitizing agents, such as oxygen. Commonly observed photodegradation products in an aerated, neutral-pH, aqueous protein solution include S-S bond fission, conversion of tyrosine to 3-(4-hydroxyphenyl)lactic acid, and dityrosine, as well as fragmentation by-products and the conversion of tryptophan residues to kynurenine and N-formyl-kynurenine (53). The position of a sensitive residue within the three-dimensional structure of the protein also influences its reactivity towards photolysis. t is also important to take into account potential damage to the protein during a lysis using circular ),UV, or fluorescent measure~entswhere incident radiation is being used.

ili

Physical instabilities, for the purpose of this chapter, have been categorizedas those reactions that do not involve the formation or destruction of covalent bonds. Aggregation of proteins, discussed primarily as a physical instability, can also occur as the result of intermolecular disulfide bond formation or transamidation as discussed previously. Conformational changes (i.e., secondary and tertiary structure changes) inproteins can occur as the result of exposure to structural perturbants or environmental stress. Often, stability studies of proteins involvechanges in temperature and ,both of which can cause structural mo ations that can be detected by , fluorescence and infrared d spectroscopic techniquessuch as spectroscopy. 3.2.1.

Aggregation and Precipitation

The term protein aggregation covers an extremelywide range of reactions and outcomes, someof which have been extensively studied for specific proteins whereas others remain challenging to study and define. The monomer form of a protein is generally the desirable physiologicallyactivespecies(with the exception of multimeric species),but there are few proteins that are notsusceptible to some form of either soluble or insoluble, reversible or irreversible aggregation. The most common mechani§m of protein aggregation is believed to involve otein denaturation and noncovalent association via hydrop~obicinterfaces. (54) (as during enaturation isusuallyinduced at gas-liquid,liquid-liquid microencapsulation processes) or container-liquid interfaces, although heat, p variations, solvents, salts, and excipients may also contribute. Thermal stress studies are often used to evaluate the rank ability of a formulation to stabilize a protein. Chan et al. showed, viadifferential scanning calorimetry SC), that calci chloride and sugars stabilized recombinant human eoxyribonuclease ( NAse) against thermal denaturation while divalent cations, urea, and guanidine hydrochloride destabilize the protein (55). A recent study on recombinant human Flt3 ligand aggregation elucidates the importance of thermal reversibility as it pertains to the mini~izationof the ligand aggregation and its potential role for determining solution conditions that can achieve the greatest long-term stability. Evidenc hat unfolding precededthe formation of the aggregate was provided by far-UV C Other noncovalent mechanisms of aggregation deionic complexation, ,and limitingsolubility salting out, charge neutralization close to the isoelectr of the molecule.Thesetypes of aggregates, unlike hy obic a ~ g r e ~ a t eshould s, be reversible, for example by dilution. Along with accelerated studies, a number of other pharmaceutical development-associated activities have beenfound to induce denaturation^ Studies on the stress of spray drying and nebulization on proteins have been published (57-60), and investigations into the effect of freeze-thaw cycling (61,62) are also documented. Formulation and fillingactivities along withshipping of liquid formulations oftensigni result teins in to gas-liquid interfaces. Surfactants such as Tween 80, stabilizeinterface-inducedaggr

but do not stabilize hG against thermal stress in SC studies. In fact, higher concentrations of the surfactants were found to destabilize the molecule (64). Formulation with sugars is also known to stabilize a number of proteins against aggregation (65,66). Soluble aggregates remain in solution often as dimer or low molecular weight multimers. In general, these types ofaggregates can be detected by high performance sizeexclusionchromatogrP-SEC) and havebeenfound by this methodin many proteins including ulin, interleukin-2 (IL-2), anti basic fibroblast growth d IFN-p (67). Conventional can also be modified to provide more accurate separation. For example, guanidine hydrochloride is often added to the mobile phase in order to differentiate native and denatured protein. In general, low levels of soluble aggregates in pharmaceutical products can be tolerated as long as the product remains stable and the soluble aggregates do notprogress to insoluble forms (68). There is concernthat increasing levels of soluble aggregate may be responsible for changes in immunogenicity of therapeutic proteins. Insolu~leaggregates usually manifest as visible particulates, although on occasion the level and sizeof the particles is so small that visual identification can be difficult. Insoluble aggregates have been documented for insulin and (69,70). Insoluble aggregates can be ted by anumber of solid-state techniques and Raman and electron spin resonance such as Fourier transform infrared ( spectroscopy. Light scattering techniques such as simple absorption at 320 and 360 nm can be useful for simple comparative evaluation otein concentration determination using IJV absorbance isoften used with a background subtraction to take into account low levels of insoluble material. 3.2.2. Adsorption

Adsorption onto various surfaces has been exploited for benefit in many disci~lines such as analytical and separation technology. There are many documented studies of adsorption in the analytical and bibmaterials literature. The propensity for proteins to adsorb is generally a problem to be overcome in the formulation and container closure selection of pharmaceutical products. Care must also be taken to evaluate parenteral administration sets for compatibility to assure that the targeted dose is delivered.Generally, adsorption is moreof an issue for low-dose therapeutics where significant portions of a dose may be lost to container/closure or administration set surfaces. In general, addition of surfactant molecules to formulations is knownto reduce adsorption losses. Calcitonin has been shown to adsorb to glass with adsorption isotherms of the Langmuir and Freundlich type, depending on pH. The addition of nonionic surfactants, such as Pluronic F68 and Tween 80,to calcitonin solutions demonstrated inhibition of adsorption and reduction of adsorption rate (71). The adsorption of proteins to liquid-gas interfaces, as well as solid-liquid interfaces, has recently become an area of interest using new techniques such as x-ray or neutron reflection. Studies haveincluded evaluation of the adsorbed amount, the total thickness of the adsorbed layer, and the effect ofpH andexcipients on these parameters (72-74). Atomic force microscopy ina force-spectroscopymode has been usedto investigate the kinetics of the adsorption process of fibrinogen mol-

ethods for Detecting Protein Degradation/Inactivation Degradation/ inactivation

Method

Deamidation Urea-IEF

Methionine oxidation

RP-HPLC RP-HPLC RP-HPLC

aFGF Hirudin Prandintide Insulin

RP-HPLC RP-HPLC RP-HPLC

G-CSF ICF- 1 ACTH

Aggregation SDS-PAGE SDS-PAGE SDS-PAGE SDS-PAGE HP-SEC HP-SEC HP-SEC UV Absorption UV Absorption AUC AUC AUC ELISA Light scattering Light scattering (dynamic) Light scattering (quasi-elastic) Viscosity PFG NMR Denaturation CD CD CD far UV CD DSC DSC DSc FTIR FTIR CZE CZE Adsorption

Peptide/ Protein

reflection X-ray Neutron reflection Neutron reflection Force-spectroscopy mode force spectroscopy

Reference Volkin et al. 1995 (78) Bischoff et a1 1992 (13) Hekman et al. 1998 (79) Ladisch and Kohlmann 1992 (80) Reubsaet et al. 1998 (22) Fransson 1997 (23) Antonini and Spoto1988 (29)

IGF- 1 human Flt3 ligand IFN-P HGH bFGF rbST sCT) bFGF Insulin bFGF vWF-A IL-2 MIP- 1a IFN-a Insulin Insulin Insulin

Geigert et al. 1988 (81) ockridge et al. 1998 (82) ~ h a h r o ket~al. 1994 (67) Chang et al. 1996 (83) Windisch et al. 1997 (84) Shahrokh et al. 1994 (67) Sluzky et al. 1991 (85) Shahrokh et al. 1994 (67) Williams et al. 1999 (86) Advant et al. 1995 (87) Varley et al. 1997 (88) Braun and Alsenz 1997 (89) Bohidar 1998 (47) Dathe et al. 1990 (90) Sluzky et al. 1991 (85)

Insulin Insulin

Bohidar 1998 (47) Lin and Larive 1995 (91)

a-crystallin Insulin IFN-y human Fit3 ligand tPA

Doss-Pepe et al. 1988 (92) Vecchio et al. 1996 (93) Beldarrain et al. 1999 (94) Remmele et al. 1999 (56)

IFN-y IL- 1 receptor Insulin IL-2 Bovine Plactoglobulin Insulin, hGH Urease Bovine serum albumin (bSA) Beta-casein Fibrinogen

Fransson 1997 (23) Remrnele et al. 1999 (56)

Radek and Castellino 1988 (95) Beldarrain et al. 1999 (94) Remmele et al. 1998 (96) Vecchio et al. 1996 (93) Prestrelski et al. 1995 (97) Rochu et al. 1999 (98) Nielsen et al. 1989 (99) Gidalevitz et al. 1999 (72) Lu et al. 1999 (73) Puff et al. 1998 (74) Hemrnerle et al. 1999 (75)

Peptide/Protein Abbreviations

Analytic eth hod Abb~eviations

ACTH -adrenocorticotropin hormone aFGF -acid fibroblast growth factor bFGF -basic fibroblast growth factor G-CSF -granulocyte colony-st~ulatingfactor IFN-interferon IGF-l -insulin-like growth factor l IL -interleukin

AUC -analytical ultracentrifugation CD -circular dichroism CZE -capillary zone electrophoresis DSC -differential scanning calorimetry t ELISA -enzyme-linked i ~ u n o s o r b e nassay FTIR -fourier transform infrared HP-SEC -high performance size exclusion chromatography IEF -isoelectric focusing PFG NMR -pulsed-field gradient nuclear magnetic resonance spectroscopy RP-HPLC -reversed-phase high performance liquid chromatography SDS-PAGE -sodium dodecyl sulfate polyacrylamide gel electrophoresis UV -ultra-violet

MIP-I a -macrophage inflammatory protein-I a rbST -recombinant bovine somatotropin sCT -salmon calcitonin vWF-A -type A domain of von Willebrand Factor

ecules onto silica (75), and various efforts at modeling adsorption have also in~estigated(76,771.

Although analytical methods use to detect and elucidate mechanisms of polypeptide degradation have been touched upon in relevant sections of this chapter, Table 4 provides a fairly comprehensive list of these methods. Typically, the evalutideswill require the useof several methods (e.g., ation of the stability of PO one method sensitive to ical changes and one method sensitive to conformational modifications).

tides and proteins are unique in the higher levels of structure that they exhibit.Loss of higher level structure typically results in lossof biological activity with or without loss to the molecule’s chemical composition.

1. 2. 3.

Clarke. Deamidation7 isomerization and racemization at asparaginyl and tides: succinimide-linked reactions that contribute to protein 262~785-794,1987. ougaard. Chemical stability of insulin 2: formation of higher ming products during storage of pharmaceutical insulin

.Deamidation of insulin during storage in frozen state. Diabetes

4. R C Strickley, BD Anderson. Solid-state stability of human insulin. I. ~ e c h a n i s mand the effect of water on the kinetics of degradation in lyophiles from pH 2-5 solutions. 3(8):1142-1153,1996. Patel, RT Borchardt. Chemical pathways of peptide degradation I. Pharm 5. Res 7(6):593, 1990.

ker, RR Bowsher, WC Mackeller, Chemical, and biological characte ic Human wth Hormone. Biotech Appl 7. Johnson, JM Shirokawa, WS mation of isoaspartate at two distinct sites during in vitro aging of human growth homone. J Biol Chem 264: 142628 . PT Wingfield, RJ Mattaliano, HF U Schemeissner. Recombinantdeamidation. Protein Eng 1:413, 1987. 9. AR Friedman, AK Ichhpurani, DM Brown, RM RA Martin, H A Zurcher-Neely, DM Cuido. Degradation of growth hormone g factor analogs in neutral aqueous solution is related to deamidation of asparagine residues. Int J Pept Protein Res 37(1): 14-20, 1991. 10. 11. 12. 31(35):8291--8299,1992. 13.

14. 15. 16. 17. 18, 19. 20. 21.

,A Van Dorsselaer, C Roitsch. Sequence-specific deamidation: isolation and biocal characterization of succinimide intermediates of recombinant hirudin. Biochemistry 32(2):725-734,1993. S Hancock, E Canova-Davis. In: oaspartate Formation in Peptides and Proteins. aton, FL: CRC Press, 1994. T Liu. Deamidation: a source of microhetero in pharmaceutical proteins. rends Biotechnol 10(10):364-369,1992. C Oliyai, RT Borchardt. Chemical pathways of peptide degradation. VI. Effect of the primary sequence on the pathways of degradation of aspartyl residuesinmodel .Pharm Res 11(5):751, 1 an, S Clarke, In: DW Aswad, midation and ~soaspa~tate Formation in Proteins. Boca Raton, Press, 1994. CT Przysiecki, J C Joyce, PM Keller, HZ Markus, CE Carty, CT Dunwiddie, R W Ellis, W J Miller, ED Lehman. Characteriza~ionof recombinant asin secreted by Saccharomyces cerevisiae.Protein Expr Purif 3(3):185-195, 1992. rti, I Ekiel, P Lindahl, M Abrahamson, AC Storer. Affinity purification and elimination of methionine oxidation in recombinant human cystatin C. Protein Expr Purif 11(1):111-118,1997. lation variables on MJ Pikal, KM Dellerman, ML Roy, RM Riggin, The effects o the stability offreeze dried human growth hormone. Pharm 4):427-436,1991. Lam, JY Yang, JL Cleland. Antioxidants for prevention recombinant monocl

22. 23.

P h a m Biomed Anal 17(2):283-289, 1998. Fransson. Oxidation of human insulin-like growth factor I in formulation studies. 3. Factorial experiments of the effects of ferric ions, E TA, and visible light on methionine oxidation and covalent aggregation in aqueous solution. J Pharm Sci 86(9): 1046-1050, 1997.

24. HS Lu, PR Fausset, LO Narhi, T Horan, K Shinagawa, G Shimamoto, T Chemicalmodification and site-directed mutagenesis of methionine re recombinant human granulocyte colony-stimulating factor: effect on stability and bio25. 26, 27. 28. 29. 30. 31. 32. 33. 34. 35.

36. 37.

38. 39. 40. 41. 42. 43.

idues in coagulation factor Katta, KR Westcott, LO Narhi, HS Lu. In vitro methionine oxidation of recombinant human leptin. Pharm Res 15(4):632-640, 1998. Manavalan, ES Cole, J es in anti thrombi^. Effects on biological activity and heparin binding. J Biol Chem 274(15):10268-10276, 199 C George-Nascimento, A Gyenes, SM Halloran, J Merryweather, P Valenzuela, asiarz, A Randolph. Characterization of recombinant human epidermal growth factor produced in yeast. Biochemistry 27(2):797-802, 1988. G Antonini, G Spoto. Reversed phase high performance liquid chromatography of adrenocorticotropin 1-39 and its fragments in the native and oxidized forms. Biochem Int 16(6):1009-1017,1988. EL Finley, J Dillon, RK Crouch, RL Schey. Identification of tryptophan oxidation products in bovine alpha-crystallin. Protein Sci 7( 11):2391-2397, 1998. WE Savige, A Fontana. Oxidation of tryptophan to oxindolylalanine by dimethyl sulfoxide-hydrochloricacid. Selective modification of tryptophan containing peptides. Int J Pept Protein Res 15(3):285-297, 1980. E Canova-Davis, J Ramachandran, Chemical modificationof the tryptophan residue in adrenocorticotropin. Biochemistry 5(4):921-927,1976. TE Creighton. Protein Structure: A Practical Approach, IRL Oxford, 1989, pp. 155-167. JY Chang. Stability of hirudin, a thrombin-specific inhibitor. The structure of alkaline-inactivated hirudin. J Biol C (17):10839-10843, 1991. ,D Soderman, MT Schaeffer, S Ortega, G DL Linemeyer, JG Gimenex-Gallego, S are neither required, present, nor compatible with full activityof human recombinant acidic fibroblast growth factor. Growth Factors 3(4):287-298,1990. D Gergel, AI Cederbaum. Inhibition of the catalytic activity of alcohol d nase by nitric oxideis associated with S nitrosylation and the release of zinc. istry 35(50):16186-16194,1996. MD Jones, J Hunt, JL Liu, SD Patterson, T Kohno, HS Lu. Determination of tumor necrosis factor binding protein disulfide structure: deviation of the fourth domain structure from the TNFR/NGFR familycysteine-richregion signature. B i o c h e ~ s t ~ 36(48):14914-14923,1997. LL Betaseron. Dev Biol Stand 96:97-104, 1998. MC Manning, P Kamlesh, RT Borchardt. Stability of protein pha~aceuticals.Pharm Res 6(11):903-918,1989. J Schultz. Meth Enzymol 11:255-263,1967. D Piszkiewicz, M Landon, El Smith. Anomalous cleavage of aspartyl-proline peptide bonds during amino acidsequence determinations, Biochem iophys Res C o r n 40(5):1173-1178,1970. C George-Nascimento, J Lowsenson,Borissenko, M Calderon, A Medina- elb by, J Kuo, S Clarke, A Randolph. Replacement of a labile aspartyl residue increasethe stability of human epidermal growth factor. Biochemistry 29(41):9584-9591, 1990. S Clarke, RC Stephenson, JD Lowenson. Liability of asparagine and aspartic acid residues in proteins and peptides. In: TJ Ahern, MC Manning, eds. Stability of Protein Pharmaceuticals: Part A. Chemical and Physical Pathways of Protein Degradation, New York: Plenum Press. 1992, pp, 1-29.

44. 45. 46, 47. 48. 49

50.

51. 52.

53. 54. 55. 56. 57.

58. 59, 60. 61. 62. 63.

nson. Parenteral formulations of proteins and peptides: stability and nt Sci Tech 42(2S):S3-S26, 1988. C Goolcharran, RT Borchardt, Kinetics of diketopoperazine formation using model peptides. J Pharm Sci 87(3):283-288, 1998. DL Higgins, MC Lamb, SL Young, DB Powers, S Anderson. The effectof the one-chain to two-chain conversion in tissue plasminogenactivator: characterization of mutations at position 275. Throm Res 57(4):527-539, 1990. B Bohidar. Light scattering and viscosity study of heat aggregation of insulin. Bio45(1):1-8,1998. ,L Langkjoer. Insulin structure and s~ability.Pharm b i o t e c ~ o l5:315-350, Nguyen, C Schoneich,RT Borchardt. Aggregation and precipitation of human relaxin induced by metal-catalyzed oxidation. Biochexnistry RB Pepinsky, M Karpusas, A Whitty, K Jones, SEGoelz. Structural and funct nglycosylated forms of human interferon-beta (IFN-beta). Pharm Res 15(4):641-649,1998. PG Burgon, PG Stanton, K Pettersson, DM Robertson. Effect of desialylation of highly purified isoforms of human luteinizing hormone on their bioactivity in vitro, radioreceptor activity and immunoactivity. Reprod Fertil Dev 9(5):501-508, 1997. R Anumula, ST Dhume. High resolution and highsensitivity methods for oligosaccharide ma~pingand characterization by normal phase performance liquid chromatography following derivatization withhighlyfluorescent a n t ~ a n i l i cacid Glycobiology 8(7):685-694,1998. DB Volkin, H Mach, CR Middaugh. Degradative covalent reactions important to protein stability. Mol Biotechnol 8(2):105--122, 1997. H Shah. Stabilization of proteins against methylene chloridelwater interface-induced denaturation and aggregation. J .Controlled Release 58(2):143-1 5 1,1999. HK Chan, KL Au-Yeung, I Conda. Effects of additives on heat denaturation of rhDNASE in solutions. Pharm Res 13(5):756-7611996. RL Remmele, SD Bhat, DH Phan, WR Gombotz. inimization of recombinant human Flt3 ligand aggregation at the Tm plateau: a mat of thermal reversibility, Biochemistry 38(16):5241-5247,1999. JD Andya, YF Maa, HR Costantino, PA Nguyen, N Dasovich, TD Sweeny, CC Hsu, SJ Shire. The effect of formulation excipients on protein stability and aerosol performance of spray-dried powders of a recombinant humanized anti-IgE monoclonal anitbody. P h a m Res 16(3):350-358, 1999. ST Tzamis, SJ Prestrelski. Activity-stability considerations of trypsinogen during spray-drying: effects of sucrose. J Pharm Sci 88(3):351-359, 1999. AY Ip, T Arakawa, H Silvers, CM Ransone, RW Niven. Stability of recombinant consensus interferon to air-jet and ultrasonic nebulization. J Pharm Sci 84(10):1210-1214, 1995. aa, PA Nguyen, SW Hsu. Spray-drying of air-liquid interface sensitive recombinant human growth hormone. J Pharm Sci 87(2):152-159, 1998. F Jameel, R Bogner, F Mauri, D Kalonia. Investigation of physiological changes to asparaginase during freeze-thaw cycling. J Pharm Pharmacol 49(5):472-477, 1997. reilgaard, LS Jones, TW Randolph, S Frokjaer, JM Flink, MC Manning, J F Carpenter. Effect of Tween 20 on free%e-thawingand agitation-induced aggregation of recombinant human factor XIII. J Pharm Sci 87(12):1597-1603, 1998. M Katakam, AK Banga, Use of poloxamer polymer to stabilize recombinant human growth hormone against various processing stresses.Pharm Dev Techno1 2(2):143-149, 1997.

ilil

64. 65. 66. 67.

68.

69.

70.

71. 72.

73.

74.

75.

lei

11, AK Banga. Effects of surfactants on the physical stability of growth hormone. J Pharm Sci 84(6):713-716, 1995. nga. Aggregation of insulin and its prevention by carbohydrate excipients. PDA J Pharm Sci Techno1 49(4): 160-165, 1995. M Katakam, AK Banga. Aggregation of proteins and its prevention by carbohydrate excipients: albumins and gamma-globulin, J Pharm Pharmacol 47(2): 103-107, 1995. Z Shahrokh, PR Stratton, GA Eberlein, YJ Wang. Approaches to analysis of aggregates and demonstrating mass balance in pharmaceutical protein (Basic Fibroblast Growth actor) formulations. J Pharm Sci 83(12):1645-1650, 1994. L Chen, T Arakawa, CF Morris, WC Kenney, CM Wells, CG Pitt. Aggregation athway of recombinant human keratinocyte growth factor and its stabilization. Pharm ng, T Calderwodd, K Tsaie, GC Visor, J Duffy, W-C Hsu, L kin-l p (IL-1p) in aqueous solution: analytical methods, kinetic products and solution formulation implications. Pharm Res 8(4):485-490, 1991. Schrier, RA Kenley, R Williams, RJ Corcoran, Y Kim, RP Nothey, D Augu uberty. Degradation pathways for recombinant human macrophage c mulating factor in aqueous solution. Pharm Res 10(7):993-944, 1993. SL Law, CL Shih. Adsorption of calcitonin to glass. Drug Dev Ind Pharm 25(2):253-256, 1999. D Gidalevitz, Z Huang, SARice. Urease and hexadecylamine-urease films at the air-water interface: an x-ray reflection and grazing incidence x-ray diffraction study. Biophys J 76(5):2792-2802, 1999. Lu, TJ Su, RK Thomas. Structural conformation of bovine serum albumin layers at air-water interface studied by neutron reflection. J Colloid Interface Sci 213(2):426-437,1999. N Puff, A Cagna, V Aguie-Beghin, R Douillard. Effect of ethanol on the structure and properties of beta-casein adsorption layers at the air/ buffer interface. J Colloid Interface n, M Maaloum, JK Horber, L Heinrich, JC Vogel, P Schaaf. irect observation of the anchoring process during the adsorption of fibrinogen on a solid surface by force-spectroscopy mode force spectroscopy. Proc Natl Acad Sci

76.

tis, S Zhang, P Rolfe. Computer modeling of the adsorpunder the influence of double layer and van der Waals Eng Comput 37(1):119-124, 1999. 77. ire, MK Bothwell. A mechanistic approach to modeling singleprotein ces. J Colloid Interface Sci 213(1):265-267, 1999. 78. DB Volkin,AMVerticelli,MW Bruner, KE Marfia, PK Tsai, MK Sardana, CR ugh. Deamidation of polyanion-stabilized acidic fibroblast growth factor. J 79.

Dixit, S Mauch, M Nuechterlein, A Stepanenko, JD d identification of peptide degradation products of heat stressed pramlinitide injection drug product. Pharm Res 15(4):650-659, 1998. 80. MR Ladisch, KL Kohlmann. Recombinant human insulin. Biotechnol Prog 8(6):469-478,1992. 81. Panschar, S Fong, HN Huston, DE Wong, DY Wong, C Taforo, M Pemberton. The long-term stability of recombinant (serine-17) human interferon-beta. ron Res 8(4):539--547, 1988. 82. ckridge, R Aston, DJ Morrell, AT Holder. Cross-linkedgrowth hormone dimers have enhanced biological activity. Eur J Endocrinol 138(4):449-459, 1998.

83, JP Chang, TH Ferguson, PA Record, DA Dickson, DE Kiehl, AS Kennington. Improved potency assay for recombinant bovine somatotropin by high-performance size-exclusion chromatography. J Chromatogr A 736(1-2):97-104, 1996. 84. V Windisch, F DeLuccia, L Duhau, F Herman, JJ Mencel, SY Tang, M Vuilhorgne. Degradation pathways of salmon calcitonin in aqueous solution. J Pharm Sci 86(3):359-364,1997. 85. V Sluzky, JA Tamada, AM Klibanov, R Langer. Kinetics of insulin aggregation in aqueous solutions upon agitation in the presence of hydrophobic surfaces, Proc Natl Acad Sci USA 88(21):9377-9381, 1991. 86. SC Williams, J Hinshelwood, SJ Perkins, RB Sim.Production and functional activity of a recombinant von Willebrand factor-A domain from human complement factor B. Biochem J 342(t3):625-632, 1999. 87. SJ Advant, EH Braswell, CV Kumar, DS Kalonia. The effect of pH and temperature on the self-association of recombinant h u a n interleukin-2 as studied by equilibrium sedimentation. Pharm Res 12(5):637-641, 1995. 88. PG Varley, AJ Brown, HC Dawkes, NR Burns. A casestudy and use of sedimentation equilibrium analytical ultracentrifugation as a tool for biopharmaceutical development. Eur Biophys J 25(5-6):437-443, 1997. 89. A Brunn, JAlsenz.Development and useof enzyme-linked immunosorbent assays (ELISA) for the detection of protein aggregates in interferon-alpha (IFN-alpha) formulations. Pharm Res 14(10):139~1400,1997. 90. M Dathe, K Gast, D Zirwer, H Welfle, B Mehlis.Insulin aggregation in solution. Int J Pept Protein Res 36(4):344-349, 1990. Lin, CK Larive. Detection of insulin aggregates with pulsed-field gradient nuclear 91. magnetic resonance spectroscopy. Anal Biochem 229(2):214-220, 1995. 92. EVV Doss-Pepe, EL Carew, J F Koretz. Studies of the denaturation patterns of bovine alpha-crystallin using an ionic denaturant, guanidine hydrochloride and a non-ionic denaturant, urea. Exp Eye Res 67(6):657-679, 1998. 93. G Vecchio, A Bossi, P Pasta, G Carrea. Fourier-t~ansforminfrared conformational study of bovine insuli~in surfactant solutions. Int J Pept Protein Res 48(2): 113-1 17, 1996. 94. A Beldarrain, JL Lopez-Lacomba, G Furrazola, D Barberia, M Cortijo. Thermal denaturation of human gamma-inte~feron.A calo~imetricand spectroscopicstudy. Biochemistry 38(24):7865-7873,1999. 95. JT Radek, FJ Castellino. A differential scanning calorimetric investigation of the domains of recombinant tissueplasminogen activator, Arch Biochem Biophys 267(2):776-786,1988. 96. RL Remmele Jr, NS Nightlinger, S Srinivasan, VVR Gombotz. Interleukin-l receptor (IL-1R) liquid formulation development usingdifferential scanning calorimetry. Pharm Res 15(2):200-208,1998. 97. SJ Prestrelski, KA Pikal, T Arakawa. Optimization of lyophilization conditions for re~ombinant human interleukin-2 by dried-state c o n f o ~ a t i o n a lanalysis using Fourier-transform infrared spectroscopy. Pharm Res 12(9):1250-1259, 1995, 98. D Rochu, G Ducret, F Ribes, S Vanin, P Masson. Capillary zone electrophoresis with optimized temperature control for studying thermal denaturation of proteins at various pH. Electrophoresis 20(7): 1586-1594, 1999. 99. RG Nielsen, GS Sittampalam, EC Rickard. Capillary zone electrophoresisof insulinand growth hormone. Anal Biochem 177( 1):20-26, 1989.

Alcon Laborato~ies(U.K.) Limited, Hemel ~ e m ~ s t e aHer~ordshire, d, England

580

1. ~ntroduction

582

easons for Stability Studies 3. Pharmaceutical Expert iscussion

of Stability

582 584

stmarketing Changes

584

evice Combinations

584 585 585

6.2. Guidelines 7.1.Impuritiesin new activesubstances 7.2. ~mpuritiesin new medicinal products 7.3. Stability testing of new drug substances and products onization guideline: stability 7.4. International Conference o testing: requirements for n 7.5. International Conference ~ ~ o t o s t a b i l itt y substances and products Committee 7.6. for nal Products guideline for *

586 587 588 589 593 594 597

7.7.Committee for Prop maximum shelf life for sterile products after first opening or following reconstitution 7.8.Committee for Proprietary edicinal Products stability guideline: abridged applications (1989 version) 8

599 599

Reprinted from Drug Development and Industrial Pharmacy, 25(7):831-856. Marcel Dekker, Inc., 1999

7.9, 7.10. 7.11.

7.12. 7.13. 7.14.

C o m ~ t t e efor Proprietory Medicinal roducts guideline: stability testing of existing activesubstances and related finished products Committee for Veterinary oducts guideline: in use stability testing of veterinary medicinal products (excluding immunological veterinary Committee for Proprietar edicinal Products guideline: inclusion of antimicrobial preservatives in medicinal products raft C o m ~ t t e for e Proprietary edicinal Products guideline: dry powder inhalers Committee for Proprietary roducts guideline: declaration of storage conditions for medicinal products in the product particulars European Union requirements for variations (changes) related

7.15. 8.

isc cuss ion

lth ~rganizationstability guidelines

600 603

604 604 605 605 608 608

Glossary

608

Appendix

613

eference

618

. I

The stability of pha~aceuticalingredients and the products containing them depends on (a) the chemical and physical properties of the materials concerned (including the excipients and container systemsused for formulated products) and (b) environmental factors such as temperature, humidity, and light and their effect on the substances in the product. ~tabilitydata on the drug substance and the f o ~ u l a t e dproduct are required in connection with applications for marketing authrorizations for pharmaceutical products containing new active substances, for products containing known active substances (including reformulations), for addiitonal strenghts or dosage forms, for new containers? and for amendments or variations to existing marketing authorizations, including new sources for active ingredients, extensions to the shelf life, and so on. In this chapter, the general re latory requirements for stability testing of pharmaceuticals for registration in the European ~ommunity(European Union) are discussedwithrespect to the sources of information on stability re~uirements,the design of stability studies for new drug substances and finished products (drug products) containing new and known (pharmacop~eialor other) active ingredients, the presentation of stability data, and conclusions. In the context of Pharmaceutical products, stability should be considered not in the dictionary definition senseof ‘‘fixed”or “notlikely to change,” but in the sense of

“controlled, documented and acceptable change.’’ The types of changethat have to be taken into account includechemical,physical, pharmacotechnical, microbiological, toxicological, as well as changes to bio-availability and other clinically significant changes. An important question then becomes: What degree of change is acceptable? This question needs to be addressed in the marketing authorization application. In arriving at an answer that is mutually acceptable to the pharmaceutical manufact~rerand the regulatory authorities, due account will need to be taken of the body of available evidence and theoretical extrapolations based on known physical and chemical lawsand statistical models. However, manyof the final proposals are often based on subjective, rather than objective, assessment of data, and the basis for decisions should therefore be discussed in the application. In addition to the need for these data, at the time of marketing authorization application it should be kept in mind that there are also requirements for stability data generation in connection with other parts of the regulatory process. This includes data generation during the conduct of safety studies under good laboratory practice conditions (e.g., the stability of the drug in the medium being used for administration in animals and the stability of the radiolabel used in tracer studies), for clinical trial approvals in manycountries (in whichboth the investigational product and any active comparator may need to be investigated for stability), as well as for preclinical and clinical studies (for which the stability of the analyte in biological specimens may need to be investigated in additional to the stability of the product under test). In addition, such studies may be needed to meet good manufacturing practice requirements (including the retention of reference samples and starting materials) after the product has been approved for marketing, to undertake stability studies on manufactured batches, and to meet the requirements of the rapid alert system with respect to products that have stability problems during distribution. Medicine inspectors may also request to see data being generated at the request of assessor, including ongoing stability investigations. Some of the types of change in pharmaceutical products that may need to be considered (depending on the dosage form concerned) are Physicalchanges: appearance, consistency, product uniformity, clarity of solution, absence of particulates, color, odor, or taste, hardness, friability, disintegration, dissolution, sedimentation and resuspendability, weight change, moisture content, particle shape and size, pH, package integrity Chemical changes: degradation product formation, loss of potency (active ingredient), loss of excipients (antimicrobial preservatives, antioxidants) icrobial changes: proliferation of microorganisms in nonsterile products, maintenance of sterility, preservative efficacy changes In designing a stability program, due account should be taken of the nature of the ingredients and the finished product, the intended clinical uses, and any special factors. The discussion in this chapter centers on consideration of the requirements that apply to pharmaceutical products containing chemical active ingredients intended for use by humans. However, stability data are required for all types of products subjected to a regulatory procedure under the terms of the EC Pharmaceutical

irectives, includinghomoeopathic products, herbal products, i ~ u n o l o g i c aprol cts, biological products, products derived from biotechnology, and veterinary medicinal products. There are a number ofspecific sources of information on the data requirements for these classes of products, but these ot discussed here. In addition, a summary of World Health 0rganization 0 ) stability study recom~endationsis included, as is a glossary of stability-related terms.

~tabilitydata areproduced to establish the storage conditions and retest interval of the active ingredient and the storage conditions and shelf life for the manufactured products. Part of the information can also be used to justify overages included in products for stability reasons. When sufficient data are not available to support fully a desired shelf life, it isopen to the manufacturer to base an application on the available information, but to offer an ongoing commitment to generate additional data and to advise the regulatory authority should any value then fall outside the agreed specifications. ecently adopted stability guidelines, the Committee for oducts (CPMP) indicates that the objective of stability t provide evidence on how the quality of an active substance or medicinal product varies with time under the influence of a variety of environmental factors such as temperature, humidity and light, and enables recommended storage conditions, re-test periods and shelf-lives to be established” (1).

The pharmaceutical Expert eport is an important tool in the regulatory submission in Europe and can be used to present a rationale and justification for the proposals made in the application, not least concerning any unusual or controversial aspects of stability data orshelf lifeproposals. In the case of a request for a shelf life based on limited experimental data, the Expert should put forward an opinion on the reasonableness of that request. In addition, the pharmaceutical Expert eport should provide a critical assessment of all other relevant data, aswell as a summary of the data in the main file, and could include a discussion of the points below. For active ingredients, a summary of the results and an analysis of the data; the variability of the stability profiles of different batches of drug substance; the most appropriate storage conditions for the active ingredient; the proposed duration of storage before retesting; the significance of the degradation products that can form on storage with cross references to the relevant information in the pharmacotoxicolo~ical Expert Report; and the reasonableness of the specification proposed for the active ingredient in relation to the available batch analysis, the analytical methods used and their validation, and the available stability data.

a discussion of the basis for the relevance of the data included in the marketin authorization application other than those relating to the intended marketing formulation and proposed marketing packaging; discussion of data relating to the potential effects of manufacturing procedures (such as heating, exposure to ionizing radiation, etc.) on the stability of the active ingredient or the finished product; a summary of the results of the stability trials and an analysis of the data, larly the assay results for the active ingredient and the content of ant degradation products; commentary on any inconsistencies in the data; a discussion of the variations in stabiity found between batches of the finishe product; an explanation of the calculations used to estimate the shelf life from the stability test data; a justification for any excesses in the quantities of active ingredients (an excipients if appropriate) in the manufactured products indicating whether they are there becasue of stability reasons or tocover manufacturi~glosses; discussion of the proposed finished product specification9 particularly a justification for the release specification limits (with cross references to propriate sections of the pharmacotoxicological data or Expert a justification for the proposed finished product storage conditions and shelf life in the light of the supporting data, particularly for the shelf life of a reconstituted or diluted product, and a commentary on the speci instructions for the preparation of the material that will be administere (particularly products administered by infusion). here should also be an interplay between the stability section and other parts of the marketing authorization application. For example, the pharmaceutical eport may need to discuss a number of items included in the developmen maceutics sectionof the application in the context of the stability data, for e compatibility of the active ingredients and the excipients usedin the formulate product and compatibility of the excipients with each other; physical parameters such as the effectof p within the specified r (including potential effects on active ingredients9 antimicrobial etc.) and any long-term effects observed in the stability studies; dissolution performance of conventional release products-the data in the development pharmaceutics and stability sections can bediscussedby the pharmaceutical Expert whenconsideringwhether a dissolution test should form part of the routine finished product test specification; a discussion of the potential for interaction, particularly between semiliqui and liquid manufactured products (including aerosols) and their containerklosure systems (sorption or leaching) and other products (such as ad~inistrationsets) with which the products must be used in clinical practice, and proposals for or justification of any proposed statements for inclusion in the summary of product characteristics or patient information leaflet; and

discussion of the potential impurities that could be present inthe active ingredient and the product, including synthetic by-products; impurities from starting materials or isomerization; residual solvents, reagents, or catalysts and degradation products and the adequacy of the analytical methods used in the stability studies with respectto their specificity, and so on. The availability (or otherwise) of synthesized referencematerials for potential degradation products and/or other impurities may need to be discussed. hen a product contains a novel pharmaceutical excipient, full data may be required as for a new active ingredient. In such cases, stability data (includi~g tests relevant to the intended performance of the novel material) will also be required for the excipient, as well as for the product containing it. The pharmaceutical Expert should discuss the data and their adequacy.

ven after an application has been approved and the product has been placedon the market, consid~rationhas to be givento the effects of changes to the manufacturing formula or manufacturing processes in the light of production experience. ioequivalence of the product and stability changes may be affected. ~ppropriate studies will therefore need to be undertaken, and applications for variations will need to be submitted as necessary. The current data requirements for such lications are discussed below.

For Europe, in the case of medical devices containing an integral substance that in other circumstances would be a medicinal product, stability data will be required as part of a file that will have to be submitted to a “notified body” and that will be referred in part to a pharmaceutical competent authority for consideration of matters relating to the “pharmaceutical” material. In the case of medical devices supplied with a pharmaceutical as a single nonreusable unit, the pharmaceutical regulatory regime will apply. In both cases, the type and extent of data thatthe regulatory agency examining the dossier will expect are likely to be similar. owever, for some types of medical device, specific adviceon stability studies maybe available in the form of European Committee for Standardization (CEN) or International Organization for Standardization (ISO) standards (for example, for contact lens care products).

The primary sources of regulatory requirements in the European Community are the pharmaceutical Regulations and irectives. In addition, there is guidanceon how to meet those requirements in a series of advisory guidelines. The relevant documents can now be accessed through the Internet. There are two primary sites of relevance. The firstis http: / /d~3.eudra.org/eudralex,which has a1 the relevant documents for human medicines that form the Rules Governing dicinal Products in the ~uropean

the European ~ g e n c yfor the Evaluation

ionssuch as the formation Associa of data requirements in the stability area, there d to be studied, with specific i~formationin the which is actually a replacement for the anne ,as well as various g~idelineson stability testing, inc

require the submission of results of ogical tests, pharmacol ~hysicochemi~al,biolo on stability and storage toxicolo~icaltests, and clinical trials. horization to support the information included in the applicati f Product Gharact ' ctive 65 /EEC). I65 rather than inclus d that stability dataon the active i ~ ~ r e d i e ~ t cation itself, it should be borne *

rmal req~irementfor all analytical thorization application to have been ed with the dossier. This includes harmac~uticsand stability sections of the file. The number of replicates of individual determinations underta~enfor lidated methods should be related to the results of the validation studies. evalidation of methods may be required when, for example, si~nificantchan to the method of man~factureof the active ingredientor changes to the com of the finished product have occurred.

Several guidelines shouldbe taken into account by those undertaking stability tests uropean marketing authorization application. The adopted a guideline on stability testing in July 198 came into effect in 1989. For new active ingredients and products containing them,

this in effect has been replaced by the guideline adopted throu that all applications submitted since 1998 will conducted according to the IC derivedguideline. y guideline (the new one r ting to requireme g active ingredientsand products containing them) was adopted with an operational date of At the same time, a gui ted on the data requiremen I variation to a market ability guideline, recent documents include a testing and another for new dosage forms ails of the guideline content is given below.

indicated above, allanalytical methods should be adequately validated. For stabit is important that the analytical methods be capable of detecting and degradation products and interaction products (active ingredient / d/or active ingredient/container components) in the finished product and should also be capable of separating these from substances unique to the active redient (e.g., §ynthetic impurities that are not degradation products). The basic re~uirernentsfor information to be included in a marketing authorization application on stability of the activeingredients and the manufactured product are su~marizedbelow. For active ingredients, S

information on the batches tested (including date and lace of manufacture, batch size, and use of ba of 1s the test methodolo normal, stress, and accelerated con~nalyticaltest procedures and their validation data, W to the assay and the determination of degradation analytical results, including initial results; conclusions; a proposed retest interval and storage conditions for the in or dosage forms, roposed speci~cation toapply at the end of the shelf life (the release specification being included elsewhere in the dossier), includin limits for impurity or degradation product content following storage under the recom~endedconditions; batches tested and the packaging used, includingdetails of the corn not those of the proposed marketing formula), packaging, batch manufacturing date and site, batch size, and use of batch; study design and methods for real-time and accelerated conditions; characteristics studied, including physical, microbiological, chemical, chromatographic, and the like characteristics of the product, as well as characteristics of the product such as potency and purity, and including otential interactions between the container and closure systemand the pro-

duct, as well as characteristics such as preservative efficacy and sterility in appropriate cases; evaluation of the test procedures, including description and validation, and consideration of potential interference in the test results; test results, includinginitial results, and information on degradation products; discussion of the available information regarding compatibility and stability with administration devices suchas infusion sets and syringes, for example; conclusions; proposed shelflife and storage conditions, including the shelflife after reconstitution or after first opening of the product, and simulated in-use stability data. Information should also be provided in the application regarding any ongoing aracteristics such as crystal form, moisture content, and particle e considered for the formulated product. These may havean effect on bioavailability, content uniformity, suspension properties, and possibly patient comfort (e.g., for ophthalmic products). Critical characteristics may needto be controlled in the finished product specification. hen a chiral substance is used in a pharmaceutical product, both the raw material and the finished product containing it should be investigated for the absence of unacceptable changes in the stereochemical purity or ratio during the proposed retest interval and shelf life. or some dosageforms, particular aspects of stability testing may need to be given particular attention. Examples includethe formation of particulates on storage in products intended for intravenous infusion; the stability of intravenous additives in the diluents proposed for use; the stability characteristics of aerosol inhalation products (especiallywhenthesehavebeen reformulated to replace chlorofluorocarbon propellants). Products in plastic containers (especially for semisolid/ liquid products, including aerosol inhalation, parenteral, and ophthalmic products) will require special consideration, particularly with respectto protection ofthe product from external factors, extraction of materials from the polymer, or release of materials from polymers (includingantioxidants, mono- and oligomers, plasticizers, catalysts, processing aids, mineral compounds such as calcium, barium, or tin, etc.); and relating to the rate of moisture vapor loss and the tightness of closures, for example.

guideline on impurities in new active substances requires characterization urities and degradation products found at or above an apparent level of 0.1% (e.g., calculated with reference to the response factor of the drug). When it is not possible to identify an impurity, a summary of the studies undertaken to try to make such identi~cationshould be includedin the dossier. Data may also be reported regarding studies undertaken on impurities present below 0.1%, although this is not normally required unless the potential impurities are expected to be unusually potent or toxic or to have pharmacological activity. However, analytical values between 0.05% and 0.09% should not be rounded up when reporting study results.

ses impurities in new medicinal products. ingredient or reaction products between ingredient, and excipients or container components?but not impurities arising from exci~ients,or extraneous co~taminants(which should not be present and should be the subject of good manufacturing practice controls), observed stability trials under the recommended storage conditions should be reported, identified, and/ or qualified when suggested thresholds are exceeded. The suggested thresholds are as follows: *

aximum daily

dose

Threshold for reporting 0.1% 0.05~0

egarding the threshold for reporting, the degradation test rocedure can be an important support tool for monitoring the manufacturing quality, as well as for use in setting a shelf lifefor the product, and the reporting level should be set below the level of the identification threshold. aximum dose daily 10mg-2 g >2 8

aximum dose daily 10mg lO-lOO~g >100 mg-2 g

>2 l3

~hresholdfor identificatio~ 1.0% or 5 pg total daily intake, whichever is lowest 0.5% or 20 pg total daily intake, whichever is lowest 0.2% or 2mg total daily inta whichever is lowest 0.1% Threshold for q u a l i ~ c a t i o ~ 1.0% or 50 ,ug total daily intake, whichever is lowest 0.5% or 200 pg total daily intake, whichever is lowest 0.2% or 2mg total daily intake, whichever is lowest 0.1%

igher thres~oldsmaybe p ed(with justification) if the target reportin it is not possible to identify particular de threshold cannot be achieved. studies undertaken should be included i r at ion products, a summary owever, these requirements will affect the design of stability studies, e choice of analytical method. tion should also be considered for pa~ticularlypotent, toxic, or pharmacolo~ically signi~cant degradation products below the state

fication for the finished product should include limits for degradation products known to occur under the stated storage conditions and should take into account the ~ualificationof those substances. e basis for the proposals should be data on the product intended to be marketed rationale should be givenfor the inclusion and exclusion of impurities from such speci~cations.

taking into account the concept of climatic regions are encompassed by climatic zones

ons and duration) should beof sufficient duration to cover subsequent use (includin~reconstitution or dilution of ropriate). The use of the same storage conditions for the finished product facilitates comparisons of the data r justified storage conditions can be included. e recommen~edstora e conditions are as follows:

~tabilitystudies con

Test

Conditions 25°C f2°C or t"Ck2"C a

ntermediate

30°C f2°C

inimum period 12 months; extended to cover retest interval orlife shelf

6 months of data from 12-month study 6 months

It would normallybe expected that datawould be generated at 25°C and at 40°C, but alternative justified conditions will be accepted, for example, in the case of products with normal storage conditions that would be lower than 25°C (toe),in which case the accelerated condition would be (t 15)”C. The intermediate condition would be used when “significant change” (i.e., failure to meet the proposed or agreed specification for an active substance) or ( a ) a 5% loss of potency from the initial batch analysis of the finished product, (b) the level of a specified degradation product the disexceeded its specification limit, (c) the product exceeded its pH limits, (d) solution performance fell outside the specification, or (e) there was a failure to meet specifications for appearance or physical properties (color, phase separation, resuspendability, deliveryper actuation, caking, hardness, etc.) occurred during accelerated testing or when a product was not suitable for testing at thattemperature (e.g., an eye ointment designed to melt below 40°C).The data at 30°C and/or 40°C can be usedto evaluate the effect of short-term excursions outside the labeled storage conditions (e.g., during shipping). Other storage conditions may be used ifjustified. Specialconsideration should be given to products that change physically or chemically at lower storage conditions, such as suspensions and emulsions(whichmay cream or sediment) or creams, oils, or semisolid preparations that may increase in viscosity. Conditions of high relative humidity are particularly applicable to solid dosage forms or other products in packs designed to offer a permanent barrier to moisture vapor loss; aqueous solutions, suspensions, and the like packaged in semipermeablecontainers may be tested under low relative humidityconditions (e.g., 10-20% RH). It is understood that, at least for the United States and Europe, the long-term test conditions for semipermeable containers should be 25°Cf2°C at 40% RH f5% RH, and accelerated conditions should be 40°C f2°C at 10-20% RH, although this is due to be reconsidered during the revision of the ICH guideline. The testing intervals should be sufficient to establish the stability characteristics of the drug substance and the dosage form. The suggested sampling frequency is every 3 months for the first year and then every 6 months for the second year and annually thereafter. The long-term studies should be undertaken in packaging that is intended for storage and distribution or material that simulates that packaging for drug substances. Containers and closure systems used in primary stability studies of finished products should be those intended for marketing. atrixing or bracketing studies may be appliedif justified. This is discussed in a separate section of this chapter (page 597). Stability tests should be designed to cover those features of the active ingredient or finished product that are susceptible to change during storage. The information collected should cover any factor that is likely to affect quality, safety, or efficacy, including chemical, physical, and microbiological characteristics. The limits of acceptability in the proposed specifications should be derived from the stability profiles obtained. The specificationwillneed to include individual and total upper limits for impurities and degradation products. Justification for the proposed levelswillneed to include reference to the levels of the impurities and degradation products that werepresent in the materials used in preclinical and clinicalstudies(i.e.,whether ornot theyhavebeen “validated”).

In addition to the accelerated and long-term studies, stress tests of the drug subshould also be used. These should be designed to determine the intrinsic ation pathways and to identify likelydegradation products and show the suitability of the analytical procedures. Three batches of drug substance, prepared at not less than pilot manufacturing scale utilizing the intended synthetic route to be used for commercial material, should be used to evaluate stability. The data obtained are used to establish a retest interval. The batches tested should also be representative of the quality of the material used in preclinical and clinical studies, as well as the intended manturing-scale material. Laboratory-scale batches can be usedfor supporting data. first three production batches should be placedin a l o n g - t e r ~study if these data are not included in the submission; the study should use the same stability protocol aracteristics expected to decrease with time,it is suggested e 95% one-sided confidence limitfor the m ptable lower specification limit be used to establish this bined if the batch-to-batch variability is smalland if sattained from statistical tests applied to the regression lines and zero time intercepts for the individual batches .g., if thep value for the level of pending on the degradation significance of rejection is more than 0.25). relationship, the data maybe transformed (e.g.,ng linear, quadratic, or cubic functions) for linear regression analysis usingarithmetic or logarithmic scales; statistical tests should be applied for the goodness of fit of the data to the assumed degradation plot. When it is not possible to combine data, the retest interval tablished according to the least stable batch. n there is little degradation or little variability, formal statistical analysis a is not usually required if visual examination is sufficient. imited extra~olationof real-time data (especially when supported by accelerated data) may be allowed. Thisshould be justified by reference to the mechanism for degradation, the goodness of fit of the statistical model, the batch sizes concerned, and available supporting data. The justification should take into account assay results and levels of degradation products and any other relevant characteristics. he storage conditions for inclusion in the labeling should be in accordance with local regulatory requirements based on a nsideration of the stability data. as ~pecificwarningsshould be included(e.g., ‘ o not freeze”).Termssuch “ambient conditions” and “room temperature” should be avoided. 7 . ~ . ~ ~ .i n i Product ~ ~ ~ d

he information on the stability of the drug substance and data on the clinical trial f o ~ u l a t i o should n be taken into account in designingstability studies for the dosa form of the finished product. The chosen design should be justified. The stability studies of the finished product should be conducted on at least three batches of the formulation proposed for marketing; the batches should be ~ a d by e a process that meaningfully simulates the intended commercial process (and ideally uses separate lots of the active ingredient) in the intended commercial

container system. WO of the batches should be at least of pilot-scale manufacture (i.e., representativ of and simulating the full-scale production process, such as a minimum of scale or 100,000 tablets or capsules). A third batch may be smaller 00-50,000 tablets or capsules). The S e quality specification should apply oposed for the commercial product. ppropriate data are not submitted atches should be placed on accelpplication, the first three producti -term stability tests using the same protocol as that reported in ata on laboratory-scale batches are not normally acceptable as primary data. They may be used to support primary data, as may information on products with related, but not identical, formulations and/or different packaging. resentation and evaluation of the data should be systematic and should hysical, pharmacotechnical (e.g., dissolution), chemical, biological, cal attrib~tes.The degree of variability of the data from the tested t the confidence that future production batches will meet the speciof their shelf life. r quantitative characteristics expected to decrease with time, it is suggested hich the 95% one-sided confidence limitfor the mean degradation e acceptable lower specification limit be used to establish this be combined if the batch-to-batch variability is smalland if satfirst obtained from statistical tests applied to the regression lines and zero timeintercepts for the individual batches .g., if the p value for the level of significance of rejection is more than 0.25). pending on the degradation relationship, the data maybe transformed (e.g.,ng linear, quadratic, or cubic functions) for linear regression analysis usingarithmetic or logarithmic scales; statistical tests should applied for the goodness offitof the data to the assumed degradation plot. en it is not possible to combine data, the retest interval tablished according to the least stable batch. n there is little degradation or little variability, formal statistical analysis a is not usually required if visual examination is sufficient. Limited extra~olationof real-time data (especially when supported by accelerated data) may be allowed.This should be justified by reference to the mechanism for degradation, the oodness offitof the statistical model, the batch sizes concerned, and available supporting data. he justification should take into account assay results and levels of degradation products and any other relevant characteristics. conditions for inclusion in the labeling should be in accor ory requirements based on a sideration of the stability ‘ not freeze’,).Termssuch as should be included (e.g., ns” and “room temperature” should be avoided. the discussion of specific uropean requirements for labeling and leaflets, later.

It is addressedto the manufacturer of the originally approved product and applies to applications for new dosage forms (i.e., a different pharmaceutical product type, including products intended for different routes of administration, e.g., oral to parenteral; and different functional types,e.g.,immediate-release tablets to modified-release tablets; and different dosage forms for the same route of administration, e.g.,capsules to tablets or solutions to suspensions) containing the sameactiveingredient as the approved product. Such products should, in principle, follow the main guideline, but it might be possible in justified cases to provide a limited amount of data at the time of submission, such as 6 months of accelerated and 6 months of long-term data from ongoing studies.

ct for studies commencing after January 1998. It is an annex to the parent The guideline is intended to be taken into account when designing light challenge tests and stress tests for marketing authorization applications for new active substances and products containing them, but it may also be used in connection with variation applications for formulation or packagingamendments.ver, photostability testing is not required as a part of the stability test pr for marketed products. The guideline does not address in-use photostability of products (but seebelow) or extend to products not covered by the parent stability guideline. hotostability tests are normally undertaken on a single representativebatch of the material under consideration, although this may need to be extended if certain changes are made to the product (e.g., to the formulation or the packaging) or depending on the data initially reported concerning photostability. The approach to the test should be systematic and involve (a)tests on the drug substance; (b)tests on the formulated product outside its immediate pack; (c) tests on the drug product in the immediate pack if there are signs of photolability in this condition, and (4 tests on the finished product in the market in^ pack if there are signs of instability in this configuration, The applicant should define and justify the limits of acceptable change, and this will determine the extent of photostability testing required. Any resulting labeling requirements are those required by the relevant regulatory authority. The light sources to be used for photostability studies and described in the eline are (a) one producing an output similar to the 65 (outdoor daylight, 10977: 1993)lID65 (indoor indirect daylight) emissi standard (such as an artificial daylight fluorescent lamp combining ultraviolet [LJV and visible outputs, xenon, or metal halide lamps) with a window glass filter for a device having a significant output below 320 nm or (b) exposure of the same sample to both a cool white fluorescent lamp designed to produce an output similar to that specified in I S 0 10977: 1993and a near-LJV fluorescent lamp having a spectral distrib~tionfrom 320 nm to 400 nm with maximum energy emission between 350 nm and 370 nm and a significant proportion of output in bands at 320-360 nm and at 360-400 nm. *

Containers for test samples should be Chemicallyinert and transparent to light. Companies shouldeither maintain an appropriate temperature control to mini~ize the potential for local effects due to temperature or include a dark control in the same environment sideby side with the test samples (unless otherwise justified). recautions should be taken to minimize changes in physical state of the samples e.g., sublimation, evaporation, or melting) and to ensure minimal interference with the exposure of the test samples. ossible interaction betweensamples and containers or protective materials should be considered and eliminated as much as possible. The study design for confirmatory testing (see below)should include an overall illumination of not less than 1.2 million lux hours and an integrated near-UV energy exposure of not less than 200 watt hours m-2 for drug and formulated product. Samples may be exposed side by side with a validated actinometric system (such as quinine for the near-UV region; the guideline includes twooptions for calibra quinine actinometers) or for an appropriate period of time using calibrated r iometers/lux meters. The results should be evaluated to determine whether acceptable levels of change are induced by lightexposure, taking into account other stability data obtained for the drug and the product concerned. Special labeling and packagin may then be identified to ensure that the product will remain within the proposed specifications for the shelf life of the product. '7.5.1.

DrugSubstanceTesting

The test program for drug substance testing should be in two phases: forceddegradation and confirmatory testing. In the forced degradation study, the drug may be tested as is or in simple solution or suspension. A, variety of ~ustified)exposure conditions may be applied depending on the photosensitivity of the drug and the intensity of the light sources, but exposure may need to be limited if extensive decompositionoccurs and maybe terminated if the material is found to be photostable. These studies are used to establish the overall light sensitivityof the material and may also be used for elucidation of the degradation pathways and validation studies. Under the conditions used for this phase of the study, degradation products may be formed that are unlikely to be formed under the confirmatory test conditions. However, the information may be useful in the development and validation of analytical procedures. If it is demonstrated that these degradation products are not formed in the confirmatory studies, they do not need to be studied further. The initial test is normallyconducted on a single batch of material; if the results are equivocal, up to an additional two batches may need to be examined. Confirmatory tests should be carried out under the conditions listed above to providenecessary information concerning the precautions needed for han~ling the drug and for its packaging and labeling. The confirmatory test is normallycarried out on a single batch of material unless the results are equivocal, which requiresan additional two batches to be examined. Testing samples should be placed in chemically inert and transparent conirect challenge studies of solid drugs mayinvolve a suitable quantity of the substance being placed as a layer not more than 3 mm deep in a glass or suitable plastic dish, protected with a suitable transparent cover if necessary.

Samples should be examined for changes in physical properties (appearance, color of solution, clarity of solution, etc.), assayed, and examined for degradation he analytical methods should have been validated for otential radation products and should be capable of resolving and dete on products, especially those appearing in the confirmatory tests uld ensure that representative portions are tested. ~ t h e rmaterials should also be homogeneous. Protected samples used as dark controls should be at the same time. e results from the forced degradation stress tests cannot be expecte establish qualitative or quantitative limits for change but should be used to develop and validate test methods for use in the confirmatory studies. The con~rmatory studies s h o u l ~identify the extent of change that occurs (and the manufacturer then has to decide whetherthis is acceptable)and will helpto identify precautions ary for the manufacture, formulation, and packaging of the finished product. from other formal stability studies shouldalso be taken into account to confi the substance will remain within justified limits at the time of use. 2.

Finished Product of formulated product should follow the sequence indicated above: fully

product, product in its primary container (unless the container is i m p e ~ e a b l eto light, e.g., a metal tube or can), and, if necessary, product in its full marketing packaging. In-use ~hotostabilitytesting may be consideredfor prodinfusion fluids and dermal products, depending on the instructions mally, only one batch of product is tested unless the results obtained al, in which case an additional two batches should be examined. es should be placed insuitable containers, taking into account the physiistics of the product and the need to ensure minimal changes of the cal state while ensuring minimal i~terferencewith the irradiation of the test les. The useof cooling or the placement of samples in sealed containers may be justified. Interactions between containers and the like and product should be considered, and any suchmaterials should be eliminated when they are not essenthe conduct of the test. roduct tested outside its primary pack should be placed in chemically inert and transparent containers. Direct challenge studies of solid products may involve a suitable quantity of the material being placedas a layer not more than 3 mm deep lass or suitable plastic dish and protection with a suitable transparent cover if necessary. Tablets, capsules, or other forms should be placed in a single layer. If direct exposure is not practical (e.g., due to oxidation), the sample should be placed in a suitable inert container (e.g., made of quartz). Product tested in the primary container should be placed to provide the most uniform exposure (e.g., horizontal or transverse placement), although special adjustments may be needed in the case of large-volume containers. es should be tested after exposure using repres~ntativesamples (e.g., of or tablets) or other samples (or solutions), ensuring homogeneous dishysical properties (appearance, clarity and color of solution, disintegration or dissolution of capsules, etc.) and chemical properties (assays for the active in~redientand degradation products) should be undertaken. Exposed and dark control samples Should be tested at the same time.

uidelinecalledReduced

Stability TestingPlan-Bracketing

and

of New Rrug Substancesand Products, is an annex to the parent ICH stabilit uideline and was issued for comment in September 1996 and adopted by th October 1997. The concept of reduced testing plans is acceptable only related substances and products such as the samedosage form (comparable specifications), same (or very closely similar) formulation, and the same product using different batches of the same active ingredient. Batch-to-batch variability should be small. A manufacturer who has used matrix or bracketing studies in developing a product should justify the chosen protocol and show that the proposed shelf life for each product strength and pack is supported using stability data on both the active ingredient and the formulated products. It is normally expected that three batches of each strength, of each container type, and of each packsize are stability tested at 25°C and 60% for 12 months and at 40°C and RH for 6 months (or alternative conditions) as per the main ICH guideline,ever, by the useof appropriate bracketing or matrixing studies, the number of samplesanalyzed at each test point maybe substantially reduced yet produce adequate information to justify the proposed shelflifeof products concerned. The manufacture and start of analysismay be sta ered. cketing or matrixing wouldnormally be acceptablein the following cases: strength changes withno or small changes inthe proportion of ingredients or achieved by varying the amount of active ingredient and one or two major excipients container size changes with the same contact materials change to an equivalent closure system change to a man~facturingsite within the same company change of batch size atrixing would normally be acceptable in the following cases: strength changes associated with significant changes to the proportions of ingredients or changes in one or two minor components changes to fill volumes in containers introduction of a closure system of nonequivalent performance changes to the manufacturing process the introduction of manufacturing by a different company 7.6.1. Bracketing

The example in Table 2 is given in the guideline for a product line having three strengths of product with the same relative composition of excipients, two types of packaging, and three pack sizes for one of these types of packaging.

gn Bracketing

Example

Dosage strength/active substance lot (A,B, C)

50 mg mg type

ack

100

75mg

~-______

=sample not tested.

n the matrising type of protocol, all product strengths, packaging materials, pack sizes, and the like are included and tested to some extent. None of the product variants are excluded from the test program. Examples of the sampling schedules for the first and second time pulls is included in the guideline and given in Tables 3 an her statistically based designs may also be used.In all cases, the manufacturer will needto ensure that the distribution of all variables is evenduring the period of the protocol. ~onsultationwith regulatory agencies may be advisable before employing complex matrix designs. Initial and final timepoints should include all samples. The statistical basis of a low testing matrix. such as a 1/3 design is better when the protocol covers several strengths and/or pack sizes rather than a product with a single strength and pack f reduction in the testing may be affected by factors such as the stability of the active ingredientand the physical stability of the dosage form. With a very stable drug in a conventional tablet formulation, a 1/3 matrix could be acceptable, but if evidenceofsignificantchangesis found during the real-time arm of the stability trial, the testing frequency will need to be increased to (for example) a 2/3 matrix for the remaining part of the study. atrix Design (First Time Point) Dosage strength/active substance lot (A,B, C) 50 rng mg

type

Pack

100 C

75 mg

First time point: forthe 1/ 3 testing matrix, 1 =sample not tested, (1) == sample tested;for the 2/3 testing matrix, 1 =sample tested, (1) =sample not tested.

esign (Second Time Point) Osage strength/active substance lot (A,B, C) 75 mg

50 mg

type

Pack

100mg

First and second time points. Test/no test convention as in note to Table 3.

revised version was released for comment in June 1997; and a considerably revised d version was adopted in January 1998, with an operational date of July 1998. e guidelineapplies to allsterile products for human useexcept radiopharmaceuticals and extemporaneously prepared or modified preparations. It is emphasized that the end user is responsible for maintaining the quality of the product to be administered to patients, but that the manufacturer (authorization holder) appropriate studies and provide relevantinformation in the product eference is also made to the European Pharmacopoeial recommendations regarding storage times and conditions of storage for particular categories of sterile products once they have been opened. The guideline suggests wording to be used for unpreserved sterile products (including injections and intravenous infusions) and for aqueous preserved products and nonaqueous products. The proposals include references to storage time (in hours or days, as appropriate) and conditions (temperature). For injectable products not containing a preservative, it issuggested that the storage time and conditions should not normally be longer than 24 h at 2" to 8°C unless reconstitution or dilution had taken place in controlled and validated aseptic conditions. For preserved sterile products, it is indicated that usage periods should not normally be more than 28 days. *

applications (1989version), it is indicated that information on the stability of the active ingredient may be determined by experimental studiesor by evidence from the scientific literature for known

substances (although comparative accelerated studies may need to be considered, for example,when there is a significant change in the route of synthesis from one approved by the regulatory authorities or when there is a significant change to the production method). Whenmultiple sources of the active ingredient are requested, stability data relating to each maybe required. ata on a minimum of two batches will be required when such studies are needed. Studies of the finished product (normally at least three batches unlessthe product is stable, i.e., no significant degradation products are found, or when the active ingredient has already been used in licensed products) will be required to establish the shelflifespecification, the shelflife, and recommended storage conditions. Specific studies will also be required when the product is labile once the container ies will berequired. These should be carried out under a variety itions: The properties of the product at te~peraturesbetween 30°C should be able to be evaluated. The meankinetic temperature e should be taken as 25°C. A variety of other storage conditions should such as below -15"C, from 2°C to 8°C (refri~eration),and freeze-thaw cycling. High humidity may be relevant (i.e.9>75% binations of elevated temperature and humidity (e.g.,40"C/75% to natural or defined artificial light conditions. The summarized (tabular or graphical) data should be presented for each batch of product. The results of ongoing studies should be provided as theybecome available, including the results from the first twoor three production batches. These results should be discussed in the Expert No labeling statement is required if the product is stable at up to 30" unlessspecial storage requirements apply (e.g., do not freeze, do not refrigerate, etc.). In other cases, a recommended temperature storage range should be stated (in "C). ata will also be required for certain variations, such as when there are major to the composition of the product, to packaging materials, or to the method of preparation of the product. Results from comparative accelerated and long-term stability studies would normally be required prior to authorization. This guidelinehas been replacedby a newer one, "StabilityTesting of Active Substances and Related Finished roduct~'~ (discussed next).

98. It came into effect in October 1998. It is intended that this guideline applyto chemically activesubstances that have been previously approved in the EU and to products containing them, but not to radiopharmaceuticals or biological/ biotechnological products. The test conditions in the guideline refer to the parent ICH guideline on stability testing of new active substances and products containing them. A systematic approach to the presentation and inter~retationof data is encouraged.

.l. Active l n ~ r ~ d i e n t s ata from stress testing and formal studies may be expected. The extent of data requirements will dependon the status of the active ingredient concerni evant (i.e., European) monographs, published data, and en a pharmacopoeial monograph is available and includes data under the rity test" and/or "transparency statement" for named degradation hen additional information on the degradation pathways is available in the published scientific literature, no stress testing will be required. Such data should be generated in other circumstances. hen a pharmacopoeial monograph is available that includes suitable degraroduct limits (but no definition of a retest interval), it would be open to the applicant to state that the material complies with the monograph immediately prior to use in the manufacture of the finished product. In this case, no stability data wouldbe required, provided that the suitability of the pharmacopoeial monograph has been established for the particular named ingredient source. Alternatively, the retest interval could be set based on results from long-term stability tests. hen data are required, the following general suggestionsare included in the guideline concerning the selection of batches for testing, although alternative test conditions can be proposed and justified: Option l: Two industrial-scale batches made using the same synthetic route and in the same packaging used for storage and distribution or one batch sim~latingthat packaging at the time of submission(accelerated and long-term study data) and a further batch (placed in a long-term test) to the same protocol after approval. Accelerated test conditions are 40°C rt 2'6, 7' of hs data. Long-term test conH; 6 months of data. ditions are 25 the same synthetic route at the Option 2: At least three pilot-scale batches using time of submission (6 months of accelerated and 12 months of long-term study data) and extended to cover the retest period and in the appropriate packaging,with the first three production-scale batches manufactured postapproval placed in a long-term test to the same protocol, with results being submitted as they become available. Results from studies should be submitted to the regulatory authorities when available. In both cases, the testing intervals should be according to the main IC guideline. Intermediate testing conditions can beusedwhensignificantchanges occur under accelerated conditions. For temperature-sensitivematerials, long-term testing should be performed at 5OCk3"C with the relative humidity conditions accelerated studies should be undertaken at 25"Crt2"C and 60% for 12 and 6 months. The usual requirements for validated analytical methods, collection of data on features susceptible to change and having a potential on quality safety or efficacy and so on apply. he specification should include limits of acceptability. This should include individual and total upper limits for impurities, including degradation products. The limits should be based on safety and efficacy considerations.

Labeling statements for active ingredients should indicate appropriate conditions based on the stability evaluation, but should avoid terms such as “ambient conditions’?or “room temperature.”

At the time the application for authorization is submitted, data will normally be expected from accelerated (6 months of data at 40°C rt 2”C/60% and l o ~ ~ - t e rstudies m of the proposed formulation and dosage for marketing. For conventional dosage forms such as immediate-release soliddosage forms and solutions and when the active substances are known to be stable (i.e., remains within initial specifications after 2 years at 25’C/60%0 months at 4OoC/75% RH), longfor 6 months at 25°C If:2”C/60% cases (e.g., prolonged-release dosage to be unstable), stability data arelikely to be requiredfor three batches, two at pilot scale and a third smaller batch, for 12 months at the same conditions. Other storage conditions may be justified, for example, if there is “significant change” under the normal conditions. The permitted conditions in special cases (e.g.,products stored under refrigerated conditions or l d products in semipermeable containers) follow guideline. The usual pattern of test sampling the advicegiven in the parent I points should be followed. In all cases, the manufacturin~method should be a simulation of that to be applied to large-scale marketing batches, and the packaging used should be the same as that used for the commercial product. The methods should investigate preservative losses, physical property changes, and, when necessary, microbiological attributes, as well as chemical and biological stability. ~reservativeefficacy tests, as well as preservative assays, should beundertaken on stored samples, particularly when different limits are proposed for preservatives in the release and shelf life specifications. In the case of particle size and/or dissolution rate specifications, reference should be made to batches used to establish bioavailability or bioequivalence. Specificationsshould be proposed for shelf lifeand for release (whenthe values differ from the shelf life specification values).The shelf life specifications, including specific upper limits for degradation products (based on efficacy and safety considerations), may deviate from the release limits provided that these are justified by the changes observed in storage. The data should be analyzed S stically to establish the degree of variability between batches (as per the parent I guideline) unless there is so little variability that it is apparent that the requested shelf life will be granted. In establishing the acceptable shelf life, limitedextrapolation of the available real-time data maybeallowed, particularly when the accelerated dataare owever, since this assumes that the degradation mechanism will continue to apply beyond the observed data points, it is important to consider what is known about the degradation mechanism and the goodness of fit of the data to any mathematical model and any other supportive data, including mass balance. It is indicated that the normal extrapolation is twice the length of the real-time studies (but not more than 3 years).

ased on the available data, appropriate storage and labeling statements should be proposed. These should refer, as appropriate, to conditions such as temperature (avoiding the terms “ambient conditions’’ and “room temperature”), light, and humidity.

Committee for Veterinary be requested for) human products by regulatory agencies. The test procedure involves removal of aliquots of the (reconstituted if appropriate) product from the container at regular intervals over the proposed in-use shelf life. he guideline isintended to offer information for the design of studiesto estabperiod for which a multiuse product may be usedafter the first dosehas been removed from the container without the integrity of the product being adversely affected, Such testing is required for all parenteral products in multidose containers, and particularly so if e headspace in the vial is replaced by an inert atmosphere during manufacture. ultidose oral or topical dosage forms should be tested if tability-related problem has been identified oncethe container has been a1 products may also need to be tested (e.g., when the active ingredient was subjected to oxidative degradation and was packed in a well-filled oxygen-impermeable container). At least two batches of product (of stated batch number and batch size; production-scale lots are preferred, although representative development-scale batcheswillbeaccepted) in the proposed marketing container-closure system should be tested. At least one of those batches should be approaching the end of its shelf life in the case of products that have chemical characteristics that will alter significantly during the shelflife(e.g., antioxidant levels).When different package sizes will bemarketed, the one offering the greatest demand on the system should be used. The test shouldsimulate as much as possible the intended use ofthe product. At appropriate intervals, aliquots of the product should be withdrawn by the usual technique, with the product being stored under the recommended conditio~s throughout the test period. In some cases, it rnay be appropriate to remove a significant proportion of the contents at the start of the test only and then to reexamine the product up tothe end of its proposed in-use shelf life; or, itrnay bemore relevant to remove aliquots on a daily basis if this is more representative of the normal use pattern for the product. More than one test may be necessary if a product may be used in different ways. In any case, the rationale for the design adopted should be explained. Chemical (active ingredient assays, antimicrobial and chemical preservative content, degradation products), physical (pH, color, clarity, closure integrity, particulate contamination, particle size), and microbiological properties (total viable

count, antimicrobial preservative efficacy, single or repeat challenge) of the product should be monitored during the study. Test procedures should be validated. The results should be summarized and tabulated. Conclusions should be made, drawing attention to any anomalous results. The in-use shelf life and any advice on discarding the product should be included in the summary of product characteristics. uitable labeling should be applied-including a space in which the user can write in the date when the first dose of product is withdrawn from the container. The in-use shelf life should also be stated on the label if space permits.

P and CVMP guideline for the inclusion of antioxidants and antimicrobial preservatives in medicinal products was released for comment in July 1996 and adopted in July 1997, and had an operational date of January 1998. to stability studies, it is suggestedthat the guideline on stability of new should be followed. Antio~idantand antimicrobial preservative lev ~uantifiedperiodically throughout the shelf lifeof the finis ative efficacy should be established using the European ology both at the end ofshelflife and at the low preservative content. Products intended for use on more than one occasi should also be tested for preservative efficacy under simulated in-use conditions. might also be necessary to investigate preservative efficacy following storage of opened or used containers during the proposed in-use shelf-life.

P guideline concerningdry powder inhalers was released for 997 and adopted in June 1998, and had an operational date 1998. It is suggested that dose, dose delivery, and particle characteri products need to be shown overthe proposed shelf lifeand over the pr of use. Active ingredient assays and determinations of related subs expected. The particular significanceof exposure to moistur guideline, with suggested testing conditions of ~5'C/60% .No special storage instructions relating to moisture exposu be required if the product complies with its shelf life specification after to such conditions. The exception to this is whenan overwrap is usedfor the but the unwrapped product was unable to withstand the suggested stability test conditions for a period of, say, 3 to 6 months (to stimulate normal use of the product). *

guideline concerning declaration of storage conditions for medicinal products in the product particulars was circulated for commentin June 1997, adopted in January 1998, and came into effect in June 1998. The guideline defines, based on the stability results, the types of statement to be included in the summary of product characteristics and the product labeling and leaflet. It is intended that the wording in the guideline be used verbatim, although additional or alternative storage statements will be allowed when necessaryto protect the product, but these statements will need to be supported with data. Any alternative storage recommendation must be achievable in practice. The use of the terms “room temperature’, and “ambient conditions” is stated to be unacceptable. When a product is stable at 25”C/60% RH and 4OoC/75% RH, no labeling statementrequired is refrigerate or freeze” may be added. nd 3OoC/60% RH, the labeling shou ucts are stable at 25” “”Do not refrigerate or freeze.” Pr not store above 30°C at 25”C/60% RH sh not store above 25°C” and may S refrigerate or freeze stable at 5”C, the label should at 2OC-8”C” and may state “Do not freeze.” Products that are stable below 0°C ate “Store in a freezer” and state a justified temperature. h regard to statements relating to protection from moisture or light, it is that the inclusion of warning statements should not be an alternative to the correct choice of container in the first place. However, the followin ments may be allowed: regarding moisture sensitivity, “Keep the container tightly closed” (for plastic bottles, etc.) or “Store in the original package” (for blisters); and regarding light sensitivity, “Store in the original container’’ or “Keep container in the outer carton.”

In the European Union, Commission Regulation 541/95 introduced a procedure for classification and assessment of applications to vary marketing aut including those related to stability. This was amended by Commissio 1 146198, which cameinto effect in June 1998. These measuresintr cept of classifying variations into three classes: minor, or t “notification” procedure); major, or type I1 (with an assessment and approval procedure); and those with changes that were so significant that they required a new marketing authorization application. The simplified type I procedure could be applied in cases for which (a) the requested changes weresupported by data generated according to the originally approved protocol and that showed that the originally agreed speci~cationswere still met, and (b) that the maximum requested shelf life wasnot more than 5 years, and that consequential changes to the product information fomed, part of the application for a variation. Examples. of such variations are discussed below.

Changes that do not meet these requirements will be subjected to the type 11 variation procedure, which includes a longer processing time and requires positive approval prior to implementation. The main types of application with stability-related changes that areidentified in the regulations as being type I (and the conditions that have to be met) are the following: 20,

Extensions to the shelflife as foreseen at the of authorization. time Conditions to be fulfilled: Stability studies have beendone according to the protocol that was approved at the time of the issue of the marketing authorization; the studies must show that the agreed end of shelf life specifications are still met; the shelf life does not exceed 5 years. 20A. Extension to the shelf life/retest period of the active substance. Condition to be fulfilled: Stability studies must show that the agreed specifications are still met. 21. Change in the shelf life after jirst opening. Condition to be fulfilled: Studies must show that the agreed end of shelf life specifications are still met. 22. Change in the shelf life after reconstitution. Condition to be fulfilled: Studies must show that the agreed end of shelf life specific~tions are still met for the reconstituted product. 23 Change in storage ~ o n d i t i o ~Condition s. to be fulfilled:Stability studies have beendone according to the protocol that was approved at the time of issue of the marketing authorization. The studies must showthat the agreed end of shelf life specifications are still met. *

Stability data will be expected to form part of the supporting package for variations for excipient replacement (providedthat the replacement isfor a material with comparable characteristics, except for vaccines and immunological products, for which a new application may be required), for coating changes on solid dosage forms (for which dissolution characteristics should be unchanged), and for changes to the container and closure system. See below for further discussion. Other variations concerneddirectlywithshelflife and in-uselife of pharmaceuticals are likely to be type I1 variations, such as those that donot comply with the conditions to be fulfilled listed above. The CPMP adopted a guideline on stability studies for type I1 variations in April 1998, and this had an operational date of October 1998. It is indicated that the guidance was intended to provide general information, but that it was intended to allowsufficientflexibility for different practical situations that maybe encountered (although alternative approaches should be scientifically justified). The guideline points out that stability studies should be continued to the approved retest interval for active substances and shelf life for products, and that the regulatory authorities should be advised for any problems that are identified. The specific examples covered in the document include changesto the manufacturing process for the active substance, a change in composition of the finished product, or a change in the immediate packaging of the finished product. In all cases, the relevant available information on the stability of the active substance (stability profile, includingstress test results, supportive data, and primary accelerated and long-term testing data) and finished product (primary data from

accelerated and long-term studies and any supportive data) should be taken into account. Appropriate additional studies will then need to be reported. The data requirements at the time of submission of the variation will depend on the particular circumstances. The applicant should investigate whether the proposed change(s) havea potential impact on the quality characteristics of the active substance and/or the product. Examples are discussed below. In addition, the first three production batches made post approval should be placedin long-term and accelerated stability studies usingthe same protocols as used for the submitted ata should be submitted to the authorities when available. Normal extrapolation of the data willbeallowed(e.g., up to twotimes the real-time data up to a limit of 3 years). 7.14.1.

Changes to the ~anufacturing Process for the Active Substance

7.14.1.1. ~ o d l ~ c a t i ~ to n ~One s ) or More Steps of the Same Route of Synthesis hen the physical characteristics, impurity profile, or other quality characteristics the active substance are adversely affected by the change, there will be a need to provide 3 months of comparative accelerated and long-term data on one batch of product at not less than pilot scale for stable drugs. In other cases, 6 months of data will be required from three pilot-scale batches. hen the change is to the final stage of synthesis (e.g., new solvent or new crystallization method), additional accelerated and long-term data (3 months, two pilot-scale batches) maybe neededfor solid dosageforms, for which the physical characteristics may have an impact on the stability of the product. 7.14.1.2.Change of SyntheticRoute Comparative accelerated and long-term data will be required on three batches of at least pilot-scale ma~ufacturefor 6 months. W e n there is a change of active substance specification that might affect the finished product (e.g., physical characteristics, degradation product levels) data may also be required on the finished product using at least two pilot-scale batches and reporting 3 months of data at the time of submission. es to the Composition of the Finished Product

For conventional dosage forms such as conventional release solid dosageforms and solutions in which the active substance is stable, there will be a requirement for 6 months of data to be supplied from twopilot-scale batches from long-term and accelerated conditions. For critical dosage forms and less stable active substances, 6 months of data from three batches of product under long-term and accelerated conditions will be required. 7.14.3.

Changestotheimmediateackaging

of theFinishedProduct

When a less protective package is introduced or there is a risk of interaction (e.g., with liquidor semiliquid products), 6 months of data will berequired from long-term and accelerated studies of three pilot-scale batches of the product.

ill

hile the WHO guidelines for stability testing are not directly applicable in the IC regions, a summary of their requirements is includedin the Appendix to this chapter. It seems probable that some account will betaken of these during the revision of the

The European regulatory agencies have adopted the ICH guidelines for stability testing, although the practical implementation may differin detail from that applied in the other IC regions. However, there is a range of other relevant guidelines-generated either by the ICH or by the C that add to the statutory re~uirementsincluded in the EtJ Directives and Regulations affecting pharmaceutical products, particularly Directives 65 /65/EEC and '75/3 18/EEC as amended by 91/507/E~C. This paper has attempted to bring together and summarizemany of the requirements and proposals included in guidance documents that affect the types of stability data required for new chemical active substances and products containing them and for existing chemical active substances and products containing them.

The t e ~ in~ this s listing are taken fromICH ~uidelinesunless otherwise indicate^.

increase the rate of chemical degradation or physical changeof an active substance or drug product by using exaggerated storage conditions as part of th al., definitive, storage programme. data, inaddition to the long term stability studies, may also be used to assess longer term chemical effects at non-accelerated conditions and to evaluate the impact of short-term excursions outside the label storage conditions such as might occurduring shipping. Results from accelerated testing studies are notalways redictive of physical change. definition similar: Accelerated stability testing.

ce; su

d with

excipients to produce the drug product. of a stability schedule so that at any time point only the samples on the extremes, for example of container size and/or dosage strengths., are tested, The design assumes that the stability of the intermediate condition samples are represented by those at the extremes. [Where a range of dosage strengths is to be tested, bracketing designs may be par~icularlyapplicable if the strengths are veryclosely related in composition (eg for a tablet range made with differentcompressionweights of a similarbasic of the same gran~lation,or a capsule range made by filling differ lug fill weights basic com~ositioninto different size capsule shells). e a range of sizes of inter-

mediate containers are to be evaluated, bracketing designs may be applicable if the material of composition of the container and the type of closure are the same throughout the range.] ncept of di~idingthe world into four zones based on defining the prevalent climatic conditions. definition similar. hotostability characteristics under ons. These studies are used to identify precautionary measures needed in manufacturing or formulation and whether light resistant packaging and/or special labelling is needed to mitigate exposure to light. type, for example tablet, capsule, solution, cream, etc, that contains a drug ingredient generally, but not necessarily, in association with excipients. ediate packaging intended for marketing. other than the drug substance in the dosage form. /labels of a drug product designatin during which a batch of the product is expected to remain within the approved shelf-life specification if stored under defined conditions, and after which it must not be used. 0 definition: The expiry date given on the individual c o ~ t a i ~ eusually r on the label) of a drug product, designates the date up to and i ~ c l u d i ~which g the product is expected to r e ~ a i nwithin ~ p e c ~ c a t i o n isf, stored correctly. It is established f o r every batch by adding the she^ life to the ~ a n u ~ a c t u r date. i~g degrade the sample deliberately. These studies, which development phase normally on the drug substance, rall photosensitivity of the material for method develgradation pathway elucidation. physical,chemical,biological, and microbiological oduct and a drug substance, covering the expected re-test period, which are claimed in the s~bmission and which will appear in the labeling. ng that is in direct contact with the drug substance or a~propriatelabel.

The combination of immediate pack and other secondary packaging such as a carton. The process of adding together the assay valuesand levels ofdegradation products to see how closely theseadd up to 100 per cent of the initial value, with due consideration of the margin of analytical precision. This concept isa useful scientific guidefor evaluating data but it is not achievable in all circumstances. The focus may instead be on assuring the specificity of the assay, the completeness of the investigation of routes of degradation, and the use, if necessary, of identification degradants as indicators of the extent of degradation via particular mechanisms. ical design of a stability schedule so that only a fraction of the total number of samples are tested at any specified samplingpoint. At a subse pling point, different setsof samples of the total number wouldbe tested. sumes that the stability of samples tested represents the stability of all samples. he differencesin the samples for the same drug product should be identified as, for example, covering different batches, different strengths, different sizes of the same container and closure and possibly in somecasesdifferent container/closure atrixing can cover reduced testing when more than one variable is being evaluated. Thus the design of the matrix will be dictated by the factors needing be covered and evaluated. This potential complexity precludes inclusionof specific details and examples, and it may be desirable to discuss design in advance with the regulatory authority, where this is possible. In every case it is essential that all batches are tested initially and at the end of the long-term testing.] hen establishingthe mean values of the temperature, the €ormulaof J ,927-929,197 1) can be used to calculate the mean kinetic temperature. ban the arithmetic mean temperature and takes into account the rrhenius equation from whichayesderivedhis formula.

e szngle test te~peraturecorresponding to the effects on c h e ~ i c a l r e a ~ t i o n ~ i n e t i ~ s d i s t r ~of ~ ueach t i o nof thefour world c l i ~ ~ tzones ic andaccordto thef o r ~ u l a d e v e l o ~Hayes e ~ ~ JD, y Journal of rmaceutical ~ciences,1971, 927-929. It is ahighervaluethan that of theaetic ~ e a tne ~ p e r a t u r e ~

of a t e ~ p e r a t u r e - t i ~ e *

substance which has not previously been registeredas a new drug substance with the national or regional authority concerned. The manufacture of either drug substance or drug product by a procedure fully representative of and stimulating that to be applied on a full manufacturing scale. For oral solid dosage forms this is generally taken to be at a minimum scale of one tenth of full production or l00000 tablets or capsules, whichever is the larger.

ance stored in the proposed packaging under storage conproposed re-test date. roduct stored in the proposed container-closure for marditions that support the proposed shelf life. 1s

valuation of experiments f o r physical, chemical, biological, biopharmaceutical and ~icrobiologicalcharacteristics of a drug, during and beyond the expected time of shelf-life and storage of samples at expected storage conditions in the intended m a r ~ e t .The results used to establish shelf-life, to confirm projected shelf-l~eand recommend storage conditions. he date when samples of the drug substance should be re-examined to ensure that material is still suitable for use. time during which the drug substance can be considered to remain within the specification and therefore acceptable for use in the manufacture of a given drug product, provided that it has been stored under the defined conditions; after this period, the batch should be re-tested for compliance with speci~cation and then used immediately. ct is expected to remain within the approved shelf-life specification provided that it is stored under the conditions defined on el in the proposed containers and closure. definition: e period oft i ~ during e which aphar~aceuticalproductis expected, if stored correctly, to comply with the spec~cation[ ‘shelf life spec~~cation: ie re~uirements to be met throug~out the shelf-life of the drug product (should not be conf~sedwith ‘release specl~cation’)J as determined by stability ~tudieson a number of batches of the product. The shelf-life is used to establish the expiry date of each batch. The combination of physical, chemical, biolQgica1 and microbiological test requirements that drug substance must meet up to its re-test date or a drug product must meet throughout its shelf life. ec The co~binationof physical, chemical, biologicaland microbiological test requirements that determine a drug product is suitable for release at the time of its manufacture.

The abilityofaph~rmaceuticalproductto retain its properties within s p e c ~ e dlimits throug~out its shelf-l~e, The followingaspects of stability are to be considered: che~ical, ~hysical, ~icrobiological and biophar~aceutical.

definition: ~tabilitytests are series of tests designed to obtain infor~ationon the stability of a ~ h a r ~ a c e u t i cproduct al in order to define its s h e ~ - l and ~ e utilisation period ~ n d e r s ~ e c ~ e d ~ a c ~and a gstorage i n g conditions.

e variation in temperature and relative humidity of storage facilities. The equipment must be capable of controlling temperature to within a range of k2"C and relative humidity to 33%. The a ual temperatures and humidities should be monitored during stability storage. ort term spikes due to opening of doors of the storage facility are accepted as U oidable. The effect of excursions due to equipment failure should be addressed by the applicant and re stability results. Excursions that exceed these ranges for more than 24 hours should be described in the study report and their impact assessed. et)

integral part of stress testin .Special test conditions for specific products (egmetereddose inhalations and creams and emulsions) may require additional stress testing. lucidate intrinsic stability characteristics. Such. ategy and is normally carried out under more r accelerated tests. ucted to provide data on forced decomposition products and decomposition mechanisms for the drug substance. he severe conditions that may be encountered during distribution and be covered by stress testing of definitive batches of the drug substance. hesestudies should establish the inherent stability characteristics of the molecule, such as the degradation pathways, and lead to identification of degradation products and hence support the suitability of the proposed analytical procedures. The detailed nature of the studies willdepend on the individual drug substance and type of drug product. This testing is likely to be carried out on a single batch of material and to include the effect of temperature in 10°C increments above the accelerated temperature test conditions (e.g., 50"C,60"C, etc); humidity where appropriate (e.g., 75% or greater); oxidation and photolysis on the drug substance plus its susceptibility to is across a wide range of pH values when in solution or suspension. sults from these studies will form an integral part of the information prothe regulatory authorities. Light testing should be an integral part of stress testing. It is recognised that some degradation pathways can be complex and that under forcing conditions decomposition products may be observed which are unlikely to be formed under accelerated and long term testing. This information may be useful in developing and validating suitable analytical methods, but it may not always benecessary to examine specificallyfor all degradation products, if it has been demonstrated that ractice these are not found.

ability data, such as stability data on early synthetic route substance, smallscale batches of materials, investigational for~ulationsnot proposed for marketing, related formulations, product presented in containers and/or closures other than those proposed for marketing, information regarding test results on containers, and other scientific rationale that support the analytical procedures, the proposed re-test period of shelf life and storage con.. S.

definition:

Supplementary data, such as stability data on small scale batches, related formulations, products presented in containers other than those proposed for ~ a r k e t i n g and other scientiJic rationa~ethat support the analytical procedures, the proposed re-test period or s h e ~ - l and ~ e storage condition^.

A period of time duringwhich a reconstitutedpreparation or the~nished dosage~orm in an opened multidose container can be used.

A provisional guideline on the stability testing of well-established drug substances in conventional dosage forms was adopted at a meeting of the Expert Committee on Specifications for Pharmaceutical Preparations in November / For products of this type, data are often available in t decomposition processes and degradation of active substances (e.g., in the /86.529) together with adequate analytical unpublished report. be concentrated on the stability of the dosage methods. In such cas form. Stability of formulated products should be adequately investigated during development of the product, including consideration of the possibility of interac between the drug substance and excipients and container system components. consideration should be givento transport and storage conditions and to the cli conditions of the places where the product is to be marketed. Appropria recommendations may be needed to ensure compliance with the shelf life for the product as indicated by the labeled expiry date. 0 ~uidelineincludes definitionsof accelerated stability testing, batch, climatic zones, expirydate, mean kinetic temperature, real-time (long-term) stability studies, shelf life, stability, stability tests, supporting stability data, and utilization period. The purposes of stability testing are identified in Table 0 guideline, and The importance of climatic zoneisemphasized in the additional data are included in the document compared with the CP A summary of the information is included in Table A2. *

Uses of Stability Studies Accelerated studies

Phase/ purpose Development: selection adequate offormulations and container-closure systems Development and dossier: determinationshelf of life and storage conditions Dossier: substantiation of claimed shelf life Quality assurance, quality control: verification that no changes have been introduced into the formulation or manufacturing process that can adversely affect the stability of the product

Real-time studies

X X

x x

Climatic Conditions Data (WHO) Measured data Climatic zone I

Air

Open data

erived Calculated Storage data room

Temperature Relative humidity

10.9"C

18.7"C

20.0"C

21°C

75%

45%

42%

45%

Temperature

17.0"C

18.7"C 25°C

21.6"C

Relative humidity

70%

52%

24.4"C

26.0"C 30°C

26.4"C

39%

54%

35%

26.5"C

28.4"C 30°C

26.7"C

77%

70%

111 Temperature

Relative humidity I'v Temperature Relative humidity

60% 52%

35%

70% 76%

Note: References to the derivation of these figures are included in the WHO document.

The recommendation in the W 0 document is for products intended for a global market to be tested under the imatic zone TV conditions, that is, real-time studies under conditions as close as possible to those that will beencountered during storage in the distribution system (with a minimum of 12 months of data being submitted with the application) and accelerated testing at 40°C ztr 2°C and 75% lthough other conditions may be used (e.g., 3 months ditions, accelerated conditions of 40°C ztr 2°C and 75% months are recommended in the case of a less stable ducts for which a limited amount of data is available. ppropriate, alternative conditions may be used, provided that acceleratedstudies are undertaken at not less than 15°C above the expected long-term storage temperature and relevant humidity conditions.

n addition to the indicated temperature and humidity conditions, it is also suggested that stability studies for liquid products shouldconsider the useof and low temperatures (e.g.,below O"C, -10°C to -20°C)freeze-thawcycles, conditions (2°C to 8°C). Light exposure mayalso be relevant for some licly available information on stability of drugs should be taken into designing studies. Intermediate test conditions (e.g., 30"Cf 2°C and 60% R beusedif significantchangesoccur(e.g., 5% loss of initial potencyvalue; any degradation product exceeding its specification limits; dissolution test failure; fail to meet requirements for physical properties or appearance) in the accelerated studies, with 6 months of data from a 1 year study being submitted with the application. Three batches of product should be tested except whenthe active ingredientis known to be fairly stable. The batches should be representative of the manufacturing process and be made at pilot scale or full scale. When possible,different batches of active ingredient should be used to manufacture the stability lots. In addition, production batches should also be tested (e.g., for stable formulations, one batch every other year; for products for which the stability profile has been established, one batch every 3 to 5 years unless a major changes has been made to the product, e.g., to the formula or to the method of manufacture). tails of the batches used inthe stability trials should be stated: batch number, &at f manufacture, batch size, pac~aging, and so on. The suggesting samplingprogram for new products is (a) real-time studies, 0, 6, 12, 24, 36 months (or more); (b) accelerated studies, 0, 1, 2, 3, (6) months (or more); except when the active ingredient is less stable or when limited data are available (when sampling every 3 months would be appropriate) in the first year, followed by sampling every 6 months in the second year and annually thereafter. For ongoing studies to support a provisional shelf life, the study may be based on samplingevery 6 months; for highly stable formulations, annual testing may be sufficient. Analytical methods should be validated, and assaysshould be stability thods used to quantify degradation products or related substances ific and of adequate sensitivity. Suitable methods should be applied to ensure that excipients remain effectiveand unchanged during the proposed shelf life. The use of a checklist to identify stability characteristics is suggested in the guideline. A stable product is onethat shows no significant degradation or changes inits physical,chemical,microbiological, and biological properties, with the product remaining within its specification. It is suggested that the stability results should be presented in a tabular or graphical format, with details of the initial and other data points for each batch of product. A standard format is suggested as an example (see Table The study report should include information on the study design, the results, and the conclusions. The stability evaluation and the derived recommendedstorage conditions and shelf liferelate to a particular formulation and method of production. Some extrapolation from the real-timedata-when supported by accelerated data-may be appropriate.

Suggested Format for Stability Test Summary Sheet ( Accelerated/real-time studies Name of drug product:

Address: Active ingredient (INN): Dosage form: Packaging: ~ a t c hnumber

Date of manufacture

Expiry date

2 3

Shelf life:

years

months

Batch size

Type batch of

(experimental,pilot plant, p r o ~ u c t i o ~ ~

2 3

Samples tested (per batch): Storage/ test conditions "C Temperature -

(lux

Light

hours)

Humidity Pressure

YORH

" . -

(bar)

RESULTS 1.Chemicalfindings: icro~iologicaland biological findings:

3. Physicalfindings: 4.

~onclusions:

Responsible

Date:

A tentative shelf life of 24 months may be proposed based on a stable active ingredient; no significant changes in a controlled stability study; similar for~ulationshaving been assigned a 24-month shelf life; a claimed shelf lifeof not more than twice the period for which real-timedata are submitted; and an undertaking on the part of the manufacturer to continue real-time stability studies until the proposed shelf life has been covered. he labeling that may be accepted may be based on storage under “normal conditions,” defined by WHO as “Storage in dry, well-ventilated premises at temperatures of 15°C-25°C or, depending on climatic conditions, up to 30°C. Extraneous odours, contamination, and intense light have to be excluded,” although alternative normal conditions may be defined locally having regard to prevalent conditions; between 2°C and 8°C under refrigeration, no freezing; low 8°C under refrigeration; store in a freezer at -5°C to -20°C; store below -18°C in a deep freezer. Additional statements such as “protect from light’’ or “store in a dry place” maybe ut not to cover up stability problems). ppropriate information may also be given for use and storage periods after the product is opened, diluted, or the like.

uir

ents listedhere

can be found in The Rules Gover~ing~ e ~ ~ c i n a l blished by the Commission of the European through the following Internet addresses: eudra.org (CPMP home page at the

egulation 541/95 of arch 10, 1995, concerning the examination of variations of the terms of a marketing authorization gran authority of a member state, as amended by Commission EC of January 26, 1965, on the appro~imationof provisions laid down by law, regulation^ or administrative action relating to medicinal products irective 75 /3 18/EEC of ay 20, 1975,on the approximation of the laws of the member states relating to analytical, pharmacotoxicological, and clinical standards, and protocols in respect of the testing of medicinal products

Quality Guidelines

Development Pharmaceutics Validation of Analytical Procedures: Methodology Stability Testing of New Active Substances and Stability Testing of Active Ingredie Stability Testing: Requirements for Photostability Testing of New Active Inclusion of Antioxidants and A Products ~ ~ ~ u r i tini eNew s Active Substances Impurities in New educed Stability Tes Guidance on Stab and Products S in the Product ParDeclaration of Storage Conditions for ticulars, Annex to Note for Guida stingNew Active of Substances;Annex to Note for Guidanc Stability Testing of Existing Active Substances and Related Finishe Maximum Shelf Lifefor Sterile Products fo or Following Reconstitution Stability Testing of Existing Active Substances and Finishe ucts Authori Stability Testing for a Type 11 Variation to a Ma ry Powder Inhalers

The Rules Governing ~ e d i c i n aProducts l in the European Union, Vol. 3, Guidelines: ~ e d i c i n aProdl uctsfor Human Use, Part A, Quality and S~otechnol~gy. ~uxembourg.European Com-

1. Committee for ProprietaryMedicinalProducts,EuropeanCommission. mission, 1998, p. 129.

University of Rhode Island, ~ i n g s t o nRhode , Island

620 620 620 62 1 62 1 622

1. Industry Trends 1.3. ~ontractingout

2.1. The ICH process 2.2. Other harmonization activities 3. Possible Increased Regulatory Concerns About the Stability of ~otanicals,Nutraceuticals, and OTCs

622 622 623 623

ability of Drugs in Channels of 4.2.Test distributed samples 4.3. The USP mailbox study 4.4.Labile products 4.5. Complaints 4.6. Returned product 5.New Approaches to Stability Testing 5.1. Constant date testing method arketplace sampling 5.3.Use of the near IR 5.4.Extensiveuseof fractional factorial designs

623 623 624 624 624 625 626 626 626 627 628 628

6. Interacting with Regulatory Agencies 6.1. Inspections Submissions 6.2.

e and regulation in stability testing functions of a guidance elopment of the FDA stability guidance 8.

Comments on somespecific aspects of the guidance 8.1. Introduction 8.2. Long-term /acce ated testing conditions . Switching toIC storage conditions Stability testing r abbreviated new drug ap~lications 8.5.Clinical trials stability (IN

eferences

629 629 630 630 631 631 63 1 63l 632 632 632 632 633 633 633 633 634 634 634 635 635 636

uring the 1990s, there was a substantial increase in the number of pharmaceuti~al companies, involved in mergers or takeovers. A s a result, a number of very large pharmaceutical companies, transnational in nature, have emerged. These companies have tremen~ousrevenues but also tremendous costs; in order to improve their bottom lines, they often give rigorous evaluation to all aspects of the operations, ~ e r g i n gsome units from different geographical locations, centralizing some activities, and contracting out to outside organi~ationssome operations that previously were totally in-house. Stability testing is not exempt from these winds of change.

At the same time as the emergence of the giant transnational pharmaceutical companies, small companies concentrating on some specialty niche that they are able to exploit rapidly because of their corporate maneuverability are becoming of or are taken over increasin~importance. Often, suchcompaniesgrowrapidly by a larger company. Also, unfortunately, not all succeed and many fall by the wayside.

As recently as the 1960s, there were many moderate-size companies that had a corporate, full-service approach. The company took complete responsibility for all aspects of the research, development, and regulatory approval of its products. owadays, there are many companiesthat contract out a variety of activities: manufacture of clinical trials materials, supervision of clinical trials, design of formulations to be marketed, preparation of regulatory submission documents, and so on. It isnow quite common for pharmaceutical companies to contract out part or all of their stability W .It is claimed that there can be substantial financial advantage to this practice wever, there are potential problems with this ractice, and it is essential, if a con house is selected to carry out stability work, onsor carefully monitor the work. a contract house for stability work should be a careful, estions to be addressed include: What is the reputation of the contract house in the pharmaceutical industry and with re~ulators? the company been providing services inthe stability area? What type of facilities e ~ u i p ~ edoes n t the com~anypossess? What is the education and training of the staff! Thus it is important to carry out a site visit to explore theseand other pertinent

e inspection of the contract house should be conducted using predetermined audit procedures. t is normally desirable that the audit team include at least two one of whom might be a consultant. he audit team should visit the contract house when work is in progress, since as well as exa~iningvarious documents it is highly desirable that there should be an o ~ ~ o r t u n i t y observe to chemists at work in the laboratory, as well as other persons involved in storage control, data management, etc. Obviously, particular atten ion should be given to the validation of stability-indicating assays. en it is decided to use the services of a contract house for stability testing, act that defines the relationship should, as far as possible, deal with all eventualities, For example, if stability data generated at the contract houseis son perceived to be of questionable value by a regulatory agency, such the contract house should, without additional charge, attend meetings at the regulatory agency and explain how the data was generated and interpreted. ven when a contract house has, after careful inspection, been selectedto perty work, its performance should be the subject of a conti~uing monitoring process.ose interested in using contract houses for stability testing ract Support Orga~izations(l), published

The results of the mergers and takeovers have been to produce a significant n ~ m b e r of transnational pharmaceutical companies that have research, developme~t, manufacturing, and marketing units in various parts of the world. In order to take advanta~eof economy of scale, there is an obvious trend for such co~paniesto try to standardize formulation design, package selection, and laboratory procedures at all their facilities throughout the world. This is not always possible.For example, a

product for distribution in a Climate Zone 4 country may require a special, heavy-duty tropical pack that is not necessary in North America, or Japan. Also, not all pharmaceutical excipients are universally acceptable throughout the world; thus, it may not be possible to have a formulation and pack that are identical in all areas. However, such matters as standard operating procedures design and implementation can be unified to a very considerable extent.

As a result of mergers or takeovers, the megapharmaceutical companies so formed are often under considerable pressure to cut costs and hence improve profitability. The size of such companies often means that a single company budget is quite insufficient as an instrument to control costs in esoteric areas, such as stability testing. There is, therefore, a tendency to develop budgets for the many individual cost centers which, together, comprise the company. In Chapter 1 of this book, I expressed my opinion about the value of stability budgets. In some companies, this activity has progressed further so that the stability costs associated with any of the company’s marketed products can be quantified. This type of data can be most useful and can provide material assistance for decisions about whether any given product of low sales is really a loss maker or a profit generator. One money-saving procedure adopted in some companies isto accelerate the use of automated and robotic equipment for stability testing. If an adequate period isallowed for the introduction and validation of this type of equipment, this approach can wellbe financially attractive, although of course initially it does require some significant capital expenditure on equipment. owever, the reduction in personnel costs will often amortize the equipment capital costs in a surprisingly short period of time. Centralization of all R&D-type stability testing in just one world-wide central laboratory can also, in some situations, reduce costs. Perhaps the most attractive method of controlling costs in the testing of market product retained samples will be an increased use of bracketing and even more sophisticated, fractional factorial designs.

International Conference on Harmonization of Technical rmaceuticals for Human Use (ICH) consists of representat pre-eminent pharmaceutical areas of the world, the European Union, the United States of America, and Japan. The stated objectives of the harmonization process are “the more economical use of human, animal and material resources and the elimination of unnecessary delay in the development and availability of new medicines while maintaining safeguards on quality, safety and efficacy and regulatory obligations to protect public health.” The ICH Steering Committee developed a five-step procedure for the development of Guidelines on a number of topics concerning research, development, and quality control of pharmaceutical products. delines most likely to be of interest to those involved in stability testing Stability Testing of NewDrugs and Products, Q2A Validation of Analytical

Although much of the ICH material is new, those expert in exegesiscan often tease out from the Guidelines the individual contributions of the Europeans, Japanese, and Americans.

Pharmacopoeias such as U.S. .,and E.P. are also involved in activities that ultimately will, it is hoped, result in considerable standardization of monographs and test methods. Such harmonization will probably take a very considerable period of time but will eventually be of great value.

At present, there is a rather obvious difference in the level of attention concerning stability issues givento prescription drugs as compared to botanicals, nutraceuticals (2), and OTCs (over-the-counter products). There is reason to believe that this difference will increasingly become a source of friction: but eventually, “harmonizati~n~~ of stability standards may prevail.

As has already been indicated in Section 3.6 of Chapter 1, our present stability testing procedures might perhaps legitimately be criticized for giving insufficient attention to the stability of “real” product in the channel of distribution, as opposed to “ideal” product kept in pampered luxury in a manufacturer’s retained stability storage area. owever, it would be quite invalid to conclude that pharmaceutical companies have not given attention to this important aspect of stability testing. Even as long ago as the 1950s9there were some companies that routinely obtained samples of their own products from the marketplace and evaluated their quality.

uring the process of developing a new pharmaceutical formulation, it is relatively common for pharmaceutical scientists to challenge products by tests specifically designed to simulate some aspect of market stress. For example, for compressed tablets a drop fragility test may be more predictive of physical stability during distribution than the more common (crushi strength) hardness testing. There is not, at present any official drop fragility test owever, one version of this test involves dropping fifty tablets, one at a time, through a 50 cm glass tube onto a steel base plate, with passing results beingobtained if no more than one tablet shows any sign of chipping.

reezelthaw tests, such as are commonly used for polypeptides and proteins are another form of market challenge tests commonly used to s i ~ u l

he practice of sending samples of a pharmaceutical product out throu number of the channels of distribution and then retrieving the samples the manufacturers' laboratories so that the effects of market evaluated has been used by some companies for man ally feasible to test all channels of distribution, ha and beta sampling errors elieve this type of study to that the results of thi of studymay wellbe, to some

ional sophistication can be added to test distri /color changelabels or temperature/ time S are fabricated so that they changecolor (often erature deemed to be critical for th uct (for example, chip will record a temperature/ti~e for the wholedistribution sequence, the information is downloaded, and hardcopy isobtained when the chip is returned to the manufacturer's laboratory. *

e c o g n i ~ i nthat ~ mail order supply of drug p important element in drug distribution, the ementation in 1996 be experienced in etween June and ous parts of the United States. The rather worr that only 84% of the packages recorded tempe The r e ~ a i ~ dwere e r subjected to temperatures above 30"C, and 21.6% experienced excessive heat (over 40°C).

that some pharmaceutical products are much more liable to ducts that areunusually sensitiveto temperat~re have been designated as labile. test for lability isbased on showing that the product ph s~ecificationseither after two weeks storage at container, at any humidity, or four weeks at 40°C in the market container, preferably at 75% relative humidity, or four weeks at 25" idity in an open dish. ere has been some misunderstand in^ concerning this definition. It does not mean that all three tests have to be performed in order for a prod~ctto be desi

is meant is that if any item fails anyone of the three above tests, it ted as (5). roposed that products whichhavebeen designated as labile with a time temperature indicator that discloses a history to 24 hours at a te~peratureof 40°C. It hasbeen reparations should have a Beyond-Use date of 120 days (i.e.9 oncedispensed or repackaged) the product shouldbeusedwithin four ommanboyina and hodes have reviewed recent trends in stability testin data concerning the effect of temperature “spikes” on assigned ata has also been presented concerning stability of compounded ars likely that in the near future more attention will be givento the status aceutical products (manufactured or compounded) during their transit through the channels of distribution and their quality when they are supplied to the final user. In particular, it likely that the stability of and also products whichhaveshipped by mailwillbesu scrutiny. Unfortunately, at resent there is a paucity of the reliable temperature/ ti to make firm conclusions about how serious the probl distribution may be. the assigned shelf lifebe ducted a computer study (7), the results of which indicate if the baseline month, it is highly unlikely that an otherwise normal product would be adversely affected by a 30°C spike of 7 hours or less.

nother source of information about how pharmaceutical products are tolerating the stress of storage and distribution after leaving the locus of production is the company complaints file. Of course, for many pharmaceutica roducts, the majority of complaints probably do not involve anyaspect of stability. wever, there is often tion of complaints which do involve, at least peripherally, some oadly speaking, stability problems that are the cause of combuted to formulation or processing (including pack selection) e producer or inappropriate storage or handling at some point d i s t r i b ~ t i o If ~ . it is found that the stability problem is due to esign, processing, or quality control evaluation of the product at the site of manufacture, then obviously the manufacturer must take appropriate hen it is found that the problem has been causedby maltreatment tion, it is not alwayseasy to determine who or what was hen causality is established, it is often difficult to take priate remedial action. f course, as with any complaints, all stability CO investigations must be documented (9).

any pharmaceutical companies havea policy that if product is still on the market at (or near) the end of the shelf lifeperiod, it can be returned, with fullcredit being given to the hospital, community pharmacy, or other entity that returned the product. Such returned product is then destroyed and the residue disposed of in an appropriate manner. However, before the returned product is destroyed, there are strong arguments in favor of a selective evaluation of some of the material, It is not recommended that every unit of returned product necessarily be subjected to a comprehensive physicochemical evaluati .In many situations it may be sufficient merelyto inspect the returned product. oes the pack still have a clean, fresh appearance? Is the label still readily legible? Is the label still adhering fully to the container? Are all seals (tamper-evident, child-resistant) still intact? Is the back-off torque of the cap still satisfactory? Does the product look, feel, and smell the same as newly manufactured material? Are there any signsof physical damage (e.g., tablet breakage)? Whenthought appropriate, specific physicochemicaltests can be applied to some samples. There is probably specific regulatory requirement that manufacturers evaluate returned product. wever, the company that is ~edicated to protecting its reputation as a producer of high-quality products has a strong incentive to do so, Since returned product was, until shortly before its return, available for supply to patients, there is good reason to conclude that its quality, when returned, is not signi~cantlydifferent from what it was whenin the marketplace. Thus although it is probably true that in many instances returned goods are not representative of the average quality of the product on the market, it is still prudent for attention to be given to monitoring the quality of returned product.

he analytical and statistical methods used in the execution and interpretation of stability studies, as well as regulatory policies, are highly unlikely to remain static. this section of the chapter, the author considers some of the possible changes stability testing that may become increasingly important in the future. It certainly cannot be guaranteed that the concepts discussed in this section will all become of importance in all areas of the WO ,nor can we be assured that there will not be other important developments.ophesyingscientific and regulatory developments is not an exact discipline!

Carstensen and Rhodes (10) proposed in 1993 the use of a stability protocol which they calledthe Constant Date method (to distinguish it from the conventional Constant Interval method). This method has attracted the attention of pharmaceutical scientists and regulators, and it appears likely that its use will become increasingly common. The conventional Constant interval method is based on the concept that the retained stability samples should be tested at fixed intervals during shelf life. Thus, if a product has a 5-year shelf life, it would normally be tested at 0, 3, 6, 9, 12,

18, 24, 36, 48, and 60 months. Once the product becomes successful with say 80 batches being manufactured per year, there will be, on average, about one stability sample requiring evaluation every other week. Setting up the assay procedure for, say, just one or two stability samples is both tedious and expensive. In the past, some companies tested some samples early and some later than the theoretical date. Unfortunately, som investigators have adopted a rather legalistic approach to the timeswhen r stability samples should be tested. In somecases, 483’s (Notice of Adversgs)havebeenissued by FRA personneleven though the testing evaluati as less than a week late. Thus there has beenincreasing rn within industry to discover a procedure that would be acceptable to the but that would not require the wasteful expense of setting up an running standards and controls, etc., all for just one or two retained ethod proposed by Carstensen and Rhodes is probably majority of pharmaceutical products. Using this approach, the stability testing of retained samples for any individual product is simply restricr specified dates each year (e.g., the first working Tuesday or February, ust, and November). Thus some samples may be tested up to about 6 weeks “early” and some up to 6 weeks “‘late,’’ but the totality of data obtained is the same (or, in SOM nstances, somewhat greater) than wou by the Constant Interval thod. Obviously, the use of the Consta must be specified in any required regulatory approval document. This method is not only more cost effective, it is also likely to improve the extent to which trend analysis of stability data is completed in a timely fashion.

hodes not only made the quite moderate proposal protocol but also made a more radical proposal etplace sample testing as a major element in the stability testing at least some pharmaceutical products. In essence, what ~arstensenand odes proposed was that if a pharmaceutical has been on the market for gnificantperiod of time (perhaps 5 years)in substantially the same form without any major stability problems, then consideration should begiven to the possibility of discontinuing, partially or completely, retained sample stability testing, relying instead on an official testing program of material sampled from the marketplace. Obviously,any marketplace testing program for the evaluation of stability must have built into it the recognition that any defect that maybediscerned in marketplace samplesmay be due to either or both of two factors: (a) a defect in the formulation design or manufacturing process, or (b) inappropriate storage or handling subsequent to the time at which the product was released onto the market by the manufacturer. Under category (b) we would, of course,includeanyadverseeffects on stability caused by a repackaging operation, Although the author is not aware of any formal official implementation of the marketplace testing of stability as proposed in ef. 10, there is undoubtedly a higher level of awareness of the importance of marketplace stress as a factor of importance governing the quality of products supplied to patients.

ar l

Near infrared spectroscopy has considerable potential in the sta hin the past decade or so, the pharmapha~aceutical products (ll). quantitative purposes has become increasceutical potential of the near I ingly evident, Now that the calibration techniques for the quantitative near I havebecomewell established and rugged,relativelyinexpensive equipment generally available. The use of this most flexible and widely a~plicablemeth of analysisispredicted to becomeincreasinglycommonin the pharmaceutical industry. In those instances where the near I spectrum of parent drug is significantly different from that of degradation products, there may be a possibility of us R to quantitate the rate of decrease of parent d g and/ or the rate of inm pecific degradation product. Since the near I can be used for nondestructive sample evaluation, this method may well prove to have considerable advant in reducing costs of stability testing. Also, this method may allowan im cision in stability testing. hen one examines the scatter seen in typical stability plots of percenta~eof label claim as a function of time, it is apparent that there are several factors that contribute to the finding that experimental data is not all n the mean regression line and thus the 90% confidence envelope is relatively wide. bviously,error derived from the limitations of the analytical method will, in many cases, be of major importance. However, for many pharmaceutical products, especially tablets, allowable content uniformity variation will play a major role in defining the variability ~ ~ envelope. of retained sample stability data and hence the width of the 9 0 confi~ence For many IJSP tablets, the allowable content u n i f o ~ i t yvariation is defined by a standard deviation of 6%. Thus evenif the analytical method had zero error and there was no other factor causing scatter on the stability plot, we would still expect to see substantial divergence between the loci of individual experimental points and the mean regression line simply because different sample tablets were, at time 0, characterized by percentage label claim values that varied perhaps from 93 to 107%. ~ommanboyinaet al. (12) have proposed the use of near stability testing of retained samples usingthe assay of the same individuallyidentified tablets for the whole testing protocol. This method removes the variance in stability data caused by content uniformity factors. As a result, it is possible that, in some cases, the improved precision of the analytical data may allow for a legitimate extension in the shelf life.

ation of some reduction in

r, thisis probably only a

rather timid beginning to a much more extensive use of well-authenticated mathematical methods that can substantially reduce the amount of stability testing with minimal effects on the levelof stability assurance provided by the experimental results (13).

Some people in the pharmaceutical industry tend to regard visits by investigators from regulatory agencies as similar to visits to a dentist-time wasting at best, remarkably unpleasant at worst. lthough this attitude is perhaps understandable on someoccasions, it ishighly sirable that, wheneverpossible, we approach visits to our facilities by investigators from regulatory agenciesin a positive manner. Our own company self-audit and other self-inspection procedures should be so com~rehensiveand rigorous in nature that we should be able to contemplate visits by investigat ulatory agencies with a quite high degree of conshould be prepared to respond to any reasonable fidence and even s comment or ques rse, unfortunately, there may occasions be when an investigator fr ory agency raises what we perceive to be a quite unreasonable iss ent of such an occurrence, we still must keep our cool, respon logical, professional, nonadversarial manner. This is not always easy, since it would be difficult to put all investigators from regulatory agencies into the category of saints. Indeed, at sometimes we might be tempted to use an epithet of quite the reverse meaning for certain investigators, owever, it is the clear responsibility of persons from the pharmaceutical industry to avoid, wheneverpossible,all personal animosity and to be cool, calm, and collected. FDA personnel that they rarely have It has been my experience in dealing with ost are hard-working individuals, simply trying to do their jobs. If we it by an investigator from a regulatory agency in a well-prepared, positive frame of mind, rather than a negative, adversarial one, we have started e company should, of course, have a well-designed, unambiguous edure) for inspections by personnel from regulatory r all aspects of the type of visit, including such matters person of the same gender throughout the visit and S during the visit or written reports left at the end of uite common for investigators to direct their questions he company escort or some seniorcompany employee. t now seems to be increasingly commonfor some investigators to ask questions of any company employee. It is therefore important thatall employees should be clearly instructed in how to respond to questions from an investigator from a regulatory

the specific points raised in the question. you don’t know the answer, say so. If, at the end of an inspection, a 483 (or some comparable document) is distributed, a well-prepared written response should be issued as soon as possible. ynes (14) has recently edited a book on F

ata pertaining to stability is often an im ortant and uite substantial part of a A. Itsuggested is that regulatory submission document, such as or some of the more important points to CO ta approval submission document are as follows. The documentshould be well prepared, easy to read, of simple linguistic style, and free of ambiguity and error. Those preparing the document should be fully conversant with all the latest information about the policies and procedures of the regulatory agency to which cation is submitted. en particularly true for stability data is the old adage that a picture can do the work of a thousand words. Thus appropriate figurescanvery often supplement tabular data and give the reader a rapid insight into the overall thrust of the results. f there are any marginal, aberrant, or unusual data, this occurrence should be commented on. It is not appropriatesimply to hope that the plethora of data will so overwhelm the reader that they will not notice a problem. This tactic might work on some occasions, but it is more likely simply to cause annoyance and delay.

here are occasions in stability work when there is no substitute for a face-to-face meeting with appropriate officials at regulatory agencies. The decisionto seek such a meeting should not be taken lightly, and the topic or topics for the meeting should ed very carefully. it is decided that a meeting is necessary, it is important to plan very carefully so that the ma~imumvalue can be obtained. In general, we should only request a meeting at a regulatory agency concerning stability matters when m is an atypical one not covered by

uidelines or other

serious and continuing personnel problem with an individual ulatory agency employee substantial chance oflegal action involving the regulatory agency It is imperative that the company clearly inform the regulatory agency in advance of the meetingof the topic or topics for discussion at the meeting, and the names and rolesof the persons whom the company plans to bring to t is often helpful to provide somesuccinctback t the meeting, the company shall present a lucid account of the problem an outline how it proposes to solve the pro m. It is normally inappropriateand imprudent to ask regulators what to do. uring the meeting, somemember of the company team, which should be kept small as possible, should take notes. ubsequent to the meeting, a brief letter should be sent to the senior official of the regulatory agencywhowas present at the meetingsummarizingwhat the company believes was the consensus reached at the meeting.

maceutical scientists involvedin stability testing obviously must be knowledgeof the principles of chemistry and pharmaceutics that underlie an understanding of the degradation of drug products. Equally true, of course, is the statement that those of us who work for the pharmaceutical industry must be aware of, and comply with, all relevant regulations and other official directives. ecause the pharmaceutical industry is pervasivelyregulated, there is a danger that some pharmaceutical scientists, working in stability testing, may become so overawed by official guidance that they relinquish all scientific judgment of their own and simply try to follow as slavishly as possible any whisper from a regulatory agency. This is most undesirable. There must never be any occasion when a regulatory guidancebecome an excusesinggoodscientific judgement. (Thecornments on guidelines provided b ean in his Packaging chapter are most appropriate to this topic.)

A guidance should be designed and used for exactly what the name implies. It should be a document that clearly defines the general principlesthat should be followedin the subject area to which it is address shouldprovidesufficient detail so that the scientist to whom it is directed sho able to understand and apply the basic principles that underlie its structure. ver, it is not possible for the authors of a guidance-although they may be most knowledgeable and experienced and possessed of considerable insight-to prepare a document that can reasonably be expected to deal with all possible problems that may arise in the field of interest. Thus, recognizing that it is not possible to forecast in advance all conceivable situations that may develop, there must be some wiggle-room in the interpretation and application of any guidance. The question then arises as to how much detail is it prudent for a guidance to provide. If we give too little information we run the risk of having more than one possibleinterpretation of how the guidance should be applied in a particular situation. If, however, we provide a plethora of detail there is a risk that our guidance may become a straightjacket, inhibiting scientific judgment and making the resolution of abnormal problems more difficult than need be. There appears to be a significantdifference in the approach used in the preparation of a guidance, between those that originate in the European and those that are developed inthe United States. The tradition in the Unite is to develop guidances that are comprehensive and quite detailed. In Europe, by contrast, guidances tend to be substantially shorter and give more emphasis to general principles than to detailed instructions.

n f Stability or of In 1987ssued the ~ u i d e l i n e f o r S u b ~ i t t i n g ~ o c u ~ e n t a t i othe This document wasof great significance in the evol~ u ~ a n and ~ioZogical~, ution of the regulation of the stability testing of pharmaceuticals. For a number of years, industry waited expectantly for a new guideline. There were delays ca I believe, primarily by the negotiations conducted under the auspices of the

In June 1998 the new, eagerly-awaited Dra ~ u b s t ~ n c and e s ~ r o d u c was t ~ issuedby the .The commentper raft Guidance was ninetyd S it was expected uld end on 31 August 199 ver, there was considerable dismay, expressed by many in the pharmaceutical industry, that the comment period days for a document comprising 3,349 lines. yielded to the arguments put to itconcerning the period allowed for As a result of public comment and extended the deadline to 31ecember1999. e to becomeofficialshowedprogressive r delays the date for the ne1998 there wer watcherswhowererelativelyconfident guidance would become official in June 1999. At the time of writing (July 1999) it now seems improbable that the guidance will become official in its final form before June 2000. Enclosed as an appendix to this book is the June 1999 uidance. The reader should be aware that the Agency has clearly label~dthe document a draft-not for implementation. er, it is this writer’s opinion, based on discussions with knowledgeable per the pharmaceutical idance are not li~ely. It is believedthat, although hly probabl~that agreement ,further delay to the issuance of the final version *

Since the guidance is generally very clearlywritten, it would be redundant to give a line-by-line exegesis.In this section attention is only directed to what may be considered to be the more salient features of the guidance.

The introduction makes it very clear that the F A guidance is based on, and fully com atible with, IC Guidelinesconcerning S bility testing of ~harmaceuticals. gives copious reference to ICH policies. A points out that applicants may chose procedures other than those e guidance, but applicants are encouraged to discuss suchaction with re departing from the procedures given in the guidance.

ive humidity values to be used the temperatures defined are it is not always easy to maintain RH values to withi ili

anufacturers of pharmaceutical products, approved by the F A for marketing in tates are placed in somewhat of a dilemma by the relatively recent

acceptance by the relevant U.S. regulatory authorities (F defined storage conditions. Such manufacturers will, in many instances, still be marketing products for which the on-going stability testing commitment was based on WO possible plans, either of

which would be acceptable which involvesa side-by-side comparison of stability under both the original and the new CH storage conditio~sis more expensive. owever, it is also the more reliable method in terms of obtaining data relevant o the effect of the change in storage conditions. Thus, if there is any doubt about which option to select it may well be prudent to select Plan

Those pharmaceutical scientists employedby generic phar likely to examine with particular attention the section on

up to24 months based on satisfactory accreted testdata at0, l , & and 3 months (for “C, the accreted conditions products to be stored at cont products notdo that meet are normally 40°C and 75% the requirements that would allow them to be classified as “simple,” the requirements are less liberal. This has caused considerable concern to some generic companies.

The guidance provides advice on the amount of stability data needed for IN trails product. In general, the required data is quite limited.

The guidance makesit clear that the normal sampling schedulefor retained stability samples is every three months during the first year, every six months during the second year, and annually thereafter. The guidance makes it clear that freezing of samples for the convenience of scheduling analysis is not acceptable. a>

racketing is given quite extensive treatment. The treatment is helpful and provides most useful advice including, for example, the statement that bracketing should not be used for clinical trials materials.

The matrixing section in the guidance is of great importance. author’s opinion that of all the sections of the guidance the discus is likely to be the one that many readers will examine more tha offers the potential of allowing a very considerable reducing in the costs associated with stability testing. Of course, the scientific principles that underlie matrixing are well established. Fractional factorial design has been used by pharmaceutical and other’ scientists for many years. The IC guidelines do allow matrixing. wever, becausethe official recognition of matrixing is so recent, there is con e uncertainty to what extent the technique maybe applied. That is wherethe guidance isso useful. It provides considerable specific information on how matrixing ance provides considerable detail, it is y matrixing stability design is put going to be considerable fle~ibilit has indicated what factors should not be matrixed. In general a matrixing design is more likely to be accepted for products with good sta~ilityand low variability rather than for those with poor stability and large variability.

A guidance makes specific r e ~ o ~ e n d a t i o nabout s thermal cycling tests. t all pharmaceutical scientistswillfind e recommendations given in the dance to be in line with current practice. any of us would recommend a 24 hour cycling period with maximum and minimum temperat~resselected as may be appropriate for the product. For many products a -4 to 40°C or 3-4 to 40°C range might be chosen. Also the practice of extending the rest to 2 or 3 weeks (i.e. 14 or 21 cycles) does seem to have merit.

guidance contains a mostusefulsection on stability testing of on, which may well have been prepared originally byseparate authors from those involved inthe other sections of the document, gives useful,and in many casesquite detailed, advice on many aspectsof the stability testing of macromolecules. The guidance frankly addr sses the several si~nificant problems that face those of us who work in this field. Lability studies of proteins can be both fascinating and fr~strating.The complex nature and hyperse~sitivity of manybiotechnology drugs make demands on pharmaceutical scientists that are often verydifficult to respond to effectively. ortunately the unit of the that deals with stability is for manymacromelecular drugs, the iologics Evaluation, and arch, is staffed by many knowle scientist/regulators, who often display considerableflexibility intheir review ofstability data for biotechnology drugs. As the guidance points out, primary data in support of the shelf life assignment for biotechnology drug must be basedon long-term, real-time conditions. Also, even

for well characterized teins it is highly unlikely that any one analytical technique willbe suf~cient toc cterizefully the stability of these products. uidance points out that for biotechnology drugs:,purity is often a relative rather than anabsolute term. Also, the guidance reminds us that itis highlydesirable to use appropriate reference materials in our stability studies. Of course, if we are developing a new chemical entity that has not previously been used as a drug:, it is highly unlikely that there will be a national or international reference standard available. It is thus extremely usefulthat the potential sponsor of a regulatory application should develop an in-house reference standard as early as possible. ince many proteins are hypersensitive to temperature stress it may not be feasible to conduct sta ty studies at temperatures other than that at which the 0, since many biotechnologydrug products are packaged product is to be stored. in glass, ~umiditystress studies may not be necessary.

normally be measured for different dosage forms. The systems to which attention is directed include Tablets Capsules ral solutions and suspensions ral powders for reconstitutions etered-dose inhalations and nasal aerosols Inhalation solutions and powders Nasal sprays: solutions and suspensions Topical, ophthalmic, and otic preparations Transdermals ~uppositories mall-volume parenterals "volume parenterals additives Implantable subdermal, vaginal, and intrauterine devices that deliver drug products The information provided in each of these sections is most useful and should be carefully examined by all those concerned with products that fall into any of the above categories.

has directed a good dealof time and energy to consideration of the amount of new stability data that should be generated when post-approval changes are made. The changes include those that involve the manufacturing process for the drug substance, manufacturing site, testing laboratory, process, site or equipment used in the fabrication of the drug delivery system or change in batch size. It seems to the author of this chapter that regulato horities in Europe and Japan have probably given much less attention than the to the implications for

stability testing that derive from post-approval changes. TheF area have been the subject of lively-in some instances quite owever even those who are presently unable to accept t roposals should be willing to commend the the sit~ationin this important area.

1. Pharmaceutical Contract Support Organiz 'on Association, Ambler, PA, 1996. 2. Kottke. Drug Development and Ind 3. nderhill, RC Rhodes 111, CT Rhodes. 25~67-70,1999. 4. CC Okeke, LC Bailey, T Medwick, LT Crady. P.F. 23:4155-4182, 1997. 5. LC Bailey, T Medwick. P.F. 245643-5644, 1998. 6. B Kommanboyina, CT Rhodes. Drug Development and ~ndustrial Pharmacy 25:857--867,1999. 7. ommanboyina, CT Rhodes. rug Development and Industrial Pharmacy, 25:1301-1306. 8. B Kommanboyina, RF Lindauer, CT Rhodes, TL Crady.International Journal of Cornpounding, accepted for publication. 9. CT Rhodes. Clin. Res. and Drug Regulatory Affairs 10:65-69, 1993. 10. JT Carstensen, CT Rhodes. Drug Development and ~ndustrialPharmacy 10:177-185,

.Pharm. Tech. (Ta~letingYearbook) 6-12,1997. Encyclopedia of Pharmaceutical Technology 19:357-370, 2000. Published by Marcel Dekker, Inc. 13. CA Lewis, D Mathieu, R Pham-Tam-Luu, eds. harmaceutical ~ ~ p e r i m e n t a l York: Marcel Dekker, 1999. Hynes, 111, ed. Preparing for FDA Pre-Approval Inspections. New York: 14,

11. 12.

J : 'GUIDANC 1707DFW.W'Pl.l 527 PR

TABLE OF CONTENTS DUCTION ..................................................... 1 LITY TESTING FOR NEW DRUG A P P L I C A ~ O N.. ~.................. 3 ~ ~ g S u b .s... ~.. ~... e........................................ 3 DrugProdu~t................................................... 7 ~ e w D o ~ g e F... o.. ~......................................... 21 OtherNT)As ................................................... 21

A. B.

D. 111.

~ R WXP.CGA O S..

. Drug sub^^ Testing

F. DataPackagefor A G. ~ t a b i lSi ~~ d Ay c A. B. c. V.

N ~DRWG ~ P L I C A ~ O ~ ILITY TESTING FOR I ~ E S T I G A ~ I ONEW P ~ ~.. e... I .................................................. 24 P ~ e .. 2..................................................... 25 P ~ e .3...................................................... 25

APPROVE^ S T A B I L P~OTOCOL ~~ ................................... 26

S ~ b iProtocol ~ i ~ ............................................... 26 Stab~lityC o ~ ~ e .. n..t........................................ 27

A. B. A.

B. C.

Vff .

.....21

........................... -22 .......................................... 22 ..................................... 22 ~ n d a ~... o.. ~........................ 22 P a ~ ~ g e ~ ~... o.. ~... e.. n... ~.. t.23 i o ~ ..... p......~......~....................... 23 ~ ... p.... ~.... ~............................. 24

OR^^ ~ T A B I DATA L ~ ....................................... 28 General ....................................................... 28 Content o f S ~ ~ l iReports ty ....................................... 28 ~ o ~ aStability ~ i Reports ~ g ....................................... 29

ECI

~ T O ~ .I... C... ~................................. 33

........................................ 33 ............................................... 35 ~ c r o b i o l o gControl i~ and ~~i~ ................................. 36 Stabili~S ~ p ~ i nC go ~ i d e . ~..~ ... o.. ~........................... 38 S ~ t i s t i ~ C o ~ iand ~ eE ~v t ~i o~ ~t .i.. o.. ~....................... - 4 1 ~ x p i ~ t i Dating o n P e ~ ~ ePeriod t e ~.............................. 44 ~ ~ c k .. e... ~... i .. ~.......................................... 46 .S

E. F.

G. Ei. I. J.

..................................................... 50 S i t e - S p ~ i Stability ~c DataForDrug and Biologic ~ ~ om ~ ..l....i...... 58 P h o t o s ~ b i l i. ~................................................. 62 Degradation Products ............................................ 71

~~~~~

. l : ‘CUIDANC 1707WFfP.WPD 5 27 98

N. N. A.

~ e ~ ~ c y c l.i.. n.. g........................................... 71 StabilityTestinginForeignLaboratoryFacilities ....................... 72 StabiIi~Testing of B i o t e c ~ o l oDrug ~ Products ....................... 73

ONS FOR SPECIFIC DOSACE F O ~ . S.................... 80 Tablets ....................................................... 80 ~ p s ues l ...................................................... 80 E m ~ i.o..................................................... 81 OraI solutio^ and S ~ ~ .... ~..... i .... o..... ~........ .-........ Oral P o w d e ~for Reconsti ................................. ~ e t e r ~ - D o I~ ~ ~ t iAerosols o ~ .........................

................................... ............................. otic P ~ ~ t .. i... O...~.................... J. .......................................... K. .................................................. L . Sml1 Volume Paenteds (SVPs) ................................... Large Volume Parentends (LVPs) .................................. DrugAdditives ......................................... rz. Implan~ble . .~U .S Vaginal and I n ~ u t e Devices ~e that ~eliver

Produc~...................................................... 85

IX. S T ~ I L TEST^^ I ~ FOR P O S ~ ~ P R CO V~~ ~ ... E.... S............ 85 A.

D. E. F.

H. I. J.

~ e n.e....................................................... 85 Process of the Drug Substance .................87 Site ...................................... 87 Change in Formulation of theDrugProduct ........................... 90 Addition o f a New Stxength for the Drug Product ....................... 92 Change in ~ a n ~ aProcess c ~ and/or g Ekluiprnentfor the Drug P r ~ u c . t ...93 Changein Batch Size o f theDrugProduct ............................ 95 R e p r o ~ ~ i naDrug ~ oProduct f ...................................... 96 Change in C o n ~ n eand r Closure of theDrugProduct ...................96 Changes in theStabilityProtocol ................................... 98

BIBLIOC~NY .......................................................... 99

~...S....~.... A.... ~............................................... 101 LIST OF T U L E S

Testing Conditions Tabk 1: Lon~-Te~Acceler&ted Table 2:

J27 96

in Drug Product Labeling

of Uniform Storage Statements

J:‘GUIDANC~1101DFIP.WPL)

................................

.............20

Table 3: ConditiQnsunder which Product has been Shown to

~ n iStorage f ~S ~ ~ t ~ r. n.. ~.. n.. ~................................ 20

....................................... 31 ....................................... 49 ....................................... 52

........................................... ................................................ le #2 ................................................ ...........................................

le #l

......................................

54 55 56 57

i Oral~Drug~P ~d u..c... ~......................................... 97

iii

ue to the length a n ~ . c o ~ Fofl this ~ i ~a~ ~ doc~~ent, ~ l e ~ e t~i ~c by line n u ~ b e r . ~ si ~~e cnco~ments

2 3 4

5 6 7

* * *

I C Harmon~ed ~ Tr~artite ~ui~eline for S t a b i l ~ T e soft iNew ~ ~ Drug ~ u b s t a ~ae s Products, Se~tember23,1994 DCH Q1A]. ICH ~ u i d e l i ~ ~ o r Testing ~ t ~ bofiNew l i ~Dosage or^ @CH QlC] ICM ~ u i ~ e l i for n e Photostabili~Te Drug Substances and Products [E ICH ~ u i ~ e l ifor n e ~ t u b i l Testing i~ l o ~ c a ~ ~ i ~ lProducts o ~ c a VCH l

11 12

13 14 15

the qualityof a drug s u b s ~ cor e ofe n ~ ~ e such n as ~ l e s ~ b l i so ~ ~ t f ~

17 18

dies &at should be p a

Investi~atio~ new

' ~ ~

ns ~ g the design, ~ conduct ~ and g use of s~bility ~to support: o ~ ~

g a p p l i ~ t i (o ~~ ~(21sCFR ) 312.23(a)(~),

p~~ bythe S~bilityT ~Cbnm ~ itccof the ~Chemistry I ~ a n u ~ ~ n g (CMEC) of the Ccntcr foe Drug Evaluation andRescad (CDER) at the Food and Controls ~ o ~ i n a t i ~i~ ng Deug Adminis~tionwith i~~ fiom the Center for Biologics valuation and (CBER). This g u i ~ ~ document representsthe Agency's cureeatthinking on stability testingof drug substancesand products. It does not create oe confer anyeights foe or on any pcrson, and does not opeeate to bind FDA oe the public. An al~mative of the applicable statute, regutations, or both. approach may be used if such an approach satisfies the requirements

~ e

0 0 0-

New drug applications (NDAs) for both new molecular entities (NMEs) and n o n - N M ~ , New dosage forms (2 1 CFR 314.50(d)( l))* A b b r e ~ anew t ~ drug a p p l i ~ t i o( ~A ~ A s()21CFR 314.92 314.99),

The

~~~

provides ~ o ~ e n ~for tthei design o ~ofstabaty studies for drug s u b s ~ c e s

these The choice oftest conditions the EU, climatic ~ n d i t i in o~ ofthe worldcan be derived fkom 28:19~202,1985, and 293947,1986). OCH QlA]

51 52

The ~ c o ~ e in nfhis ~guidance ~ oare~effective upon publi~tionof the final ~i~~ should be followed in preparing new a p p I i ~ t i or~ ,~ u b ~and ~ supplements. i o ~ ~ This represents FDA's cment thinkhg onhowthe ~ b i l i t ysection of biologi~ appli~tio~ should be ~~~. An applicant may choose to useal chooses to depart fiom the ~ ~ e n d a tseti forth o ~in this enco~ged to discussthe ma- with FDA prior to ~itiatingstudies that may later bed e t e ~ ~ to be ~ ~ p ~ b l e .

55 56 57 58

that the time necessary for applicants to establish new procedures, install, and ionthe new temperature and relativeh ~ d i ~ ~ n ~ l l ~ carry~ out o m ~ ~ b ~ e appropriate ~ b i lstudies i ~ on batchesof product, and submit thei ~ o ~ a t i in on an appli~tion may prevent some applicants from ~ e n e data ~ ~consistent g with the r e ~ ~ e n ~in the t i o ~ ~ and n g guidance for some time. However, since this guidance representsFDA's current ~ r e c o ~ e ~ d a tregarding io~ stability, submissionof data not ~nformimgwith this guidance is J :*GtJtDAhCI707DFW.WPD S 27 98

possiblewith j ~ t i ~ ~ t i o n . A p p l i ~ t i o ~ priorwtoi publication ~ d r a ~ of this guidance shou~d

5 6

72 73

74 75 76

T7 7

79 0 81..

87 8 89 90 91 92 93 94

some d e ~ d a t i o n p a t hcan ~ y be s complex &d that unde ucts m y be observed that are unlikelyto be formed under

o n be useW in developing andv ~ l o n g - ~ e ~ t eThis ~ n g~ , o ~ t i may ,but it may not always be n e c e s to ~ e ~ i n specifically e for all deg en d e ~ o ~that ~ in t practice ~ d these are not formed.

ility studies are tend^ to show that a drug s u b s ~ c will e remain within s ~ c i ~ ~ t iduring o n s the retest period if stored under~ e c o ~ e n storage d ~ d c o n d i t i ~ piCH ~.

Q W

l . Selection of Batches

tun of 12 m o n ~ '

1 1 I

~ u ~ont iatol n

1 l 1IO 111 112

2. Test P

113 114

r and Test~C ~ t e ~~ a

results of v ~ i ~ studies. ~ o n@CH QlAl

l15 1 l6

Limitsofa ~ ~ ~ b should i l ibe~de~vedfrom the

l i i t s for ~ p ~and deg~dation t i ~ p ~ u cthe~ , thelevelsobserved in materialused in ~~1~~ 4. Storage ~ o n d i t i o ~

The the studies and

the storage c o n d i ~ should o ~ be S

The t ~ ~ p r o d u c t i obatch n and ~ a n u f ~ c g u r i n ~ s c a ~ e ~ r bufch ~ u c tai omnused i n t e ~ ~ g ~ b l y ~ u g h o u t this guidance to meana batchof drug subs~nceor drug product m ~ u f a c t uatr ~ the scale ~ p i ~ l en~untcrcd ly in a facility intended for ~ a r k e p~duct;on. ~ ~ ~ g 4

29 30

ative h ~ i d conditio^ i ~ for that t e m p e ~ t ~ eThe ) . designated long-term testingcondit~o 11be r e ~ ~ t in ed the l a ~ l i n gand retest date.VCW QlA]

131 132 133 1

1 1

137

tested as soon as the a ~ e i e ~study t e ~shows s

~c h ~~g in e~ the t

In such c a s e s > ~ t e ~ aapproaches, te such8s q

~ i higher ~ ~ g ts teom ~ gpeor ,e ~ ~ sure, should beco~idered during drug develop~en~

149 150 151

ant, shorter retestp e ~ o dr, e ~ g ~ ~ t o r

154

m testing atan inte~ediatestorage ~ndition)may excursions outsidetke l a ~storage l ~nditionssuch

155 156 157

%he equip men^ must be capable of cont~lling t e r n ~ ~tor aerange of *2 "Cand relative humidity to i 5% e m ~ t u rhumiditi~ ~ a should d bemonito~during stabiiitystorage. S h o r t - t ~spikes dueto opening of doors of the storage facility are accepted as ~a~oidable. "he effect o ~ e x c u r s i due o ~ to ~ u i p mf~ t ~ 1 should be addressed by the applicant reported and ifjudged to impact stability results. Excursions that ex ranges (i.e., i 2"C andlor i 5% RH)for more than24 hours shouldbe described and their impact assessed in study the report.

RH.The actual ~

I: !GUIIIANC I707QFX3.WPD 5 27 98

159 160 161

Frequency of testing should be sufficient to establish the stabilityc ~ c t e ~ s t iofc the s drug substance. Testing under the definedl o n g - t e ~conditions will n o ~ l l be y every 3 m o n over ~ the first year, every6 months over the second year, and then ~ ~ @C 1 ~ .

6. P a c ~ g i n g/Containers

162

sho~~ The c o n ~ tonbe~used in the long-term, real-time stabili~e v ~ ~ t i o n be the simulate the actual ~ c ~ used g for~storage g and ~ s ~ b u t i o ~

7. E v ~ ~ ~ o n

gradation ~ l a t i o will ~~p analysis. U s ~ the y relati

The ckta m y show so little d e ~ ~ ~ tand i oson 1' e to go through ~ theformal s stathtical ~ ~ ~ y s ips~; v i ~ anfull g j omission is usually dlcient. gCH QlA] ~ 191 192 193 19~ 196

~ for the ~

Limited e ~ ~ l a t i may o nbe of the r e ~ - data ~ ebeyond the observ extend retest period at approval time, particularly where thea ~ l e r data a ~ ~ o w e vthis ~ , assumes that the samed e g ~ a t i o n~ l a t i o will ~ ~~p n ~ tou e observed data, and hence the useof e ~ ~ ~ tmust i obejustified n in each appli~tionin t e r n of what is ~o~ about such factors as the mechanismof degradation, the goodnessof fit of my ~ t h e ~model, t i batch ~ si=, and existence of s u p ~ ~data. v e Any eval~tionshould cover 6

197 19

not only the assay, butthe

Q W

levels of degradation products andother approp~atea~butes.

199 200

207 1. 09 210 211 212 213 14 215 216 217 218

30 tible to change d ~s t o ~ gand eg * test ~ r ~ should ~ be ~ hily e valis and the assays should be stabiii~-~ndicat~g. The need for replication will depend on the ~ s u l ~ 7

234

of validation studies.DCH Q1 A]

236

The range of testing should cover not only chemical and biological s ~ b i l ibut ~, preservative, physical properties andc ~ a c t e ~ s t i c s , o r g ~ o l eproperties, ptic and where required, ~ i ~ ~ b i oa ~ bou tge si . ~ P~~ s e ~ aefficacy t i v e testing and assays on stored samples shod^ ed out to determine the content and e ~ c a c yof ~ t i m i c ~prese~a~ves. bi~ ~C~~~A~

39 40 ~ 4 1 242

4. S ~ i f i ~ t i o ~

acceptable andj

~deviations ~ l h em the release

248 249

5. Storage Test Conditions 25 1 2 253 254 255 256 257

263

).see times. Assurance that long-term testing willc o n ~ u to e cover the e provided. PCH Q l A]

cshelf t life ~ shouid be

~ o n g - t storage e~ tem t may i oshow ~ an~ emdsions Which may sediment, orcream, oils and s e ~ - s o l i d~ ~ p ~which is accelerated increased viscosity). W e r e a lower t e r n ~ ~ ~ e c used, o n ~the~ 6-month o n testing s h o ~ be d carried out at ta e m p e at~ 1~ storage t e ~ (together ~ ~ witha~p p ~ ep ~ arelative te ~~i~~ ~ n ~ tfori that o t~ ~ ~ For example, fora product to be storedl o n g - t e ~under ~ ~ ~ e r a conditio^, ted acceIe~t~ testing should be conducted at 25°C f 2"C/60% W f 5%. The desi ~ n d i t i will o ~ be reflected in the l a b e ~ and ~ g e ~ - ~da ~te.oDCIi n QlA] Storage under~nditionsof high relativeh ~ i dapplies i ~ p ~ i c u l ~tol ysolid dosageforms, For i o ~ in packs designed to provide a dmg products suchas solutions and s u ~ e ~contained e anent barrier to water loss, specific storage under conditionsof high relative h~~~ is not necessary but the same range o f t e m p e ~ should ~ s be applied. Low relative humidity (e.g., 10 20% RH) can adverselya f k c t products packed in s e m i - ~ ~ ~ b l e c o (e.g., n ~ solutio^ ~ e r s in

,nose drops in small plastic c o n ~ n e ~and ) , consideration should be given to appropriate testing under such conditions.(ICH Q1A]

272 273 274

275 6

6 Months

Whex s i g n ~ clzunge c ~ ~occurs due to accelerated testing, additional testing atan i n t e ~ e d i ~ n d i t i o n(e.g., 3 O ' C t 2 ' ~ W 6 ~rt 5%) ~ should be conducted!, ~ i ~ ~ cchQnge a n tat d conditionsis defined as:

1. A 5percent potency loss fkom the initial assayvalue of a batch. 2. exceeding its de tion on limit. 3. The product exceeding its pH limits.

281 282 283

tuation, caking,hardness)(ICH QlA].

285

change occur at 40°W5%RH, the initial application should include a

287 289 290

X€ my p

efaibt ~ ~~ g ~change ~ c criteria a n ~during the accelerated stability study, testing of

292 293

product. ~ ~ r n a ~ v ethe, l ystudy , at the~ t e ~ ~ con~tion i a t e would bestarted from the hitid time point. a s i ~ ~ c Qchange n t o w w during12months of storageat 3 ~ ~ 6 0 appropriate to label the drug product for CRT storage with the proposed d stability data &m the full long-term studies at25'~60%W ,a l t ~ approaches, e suchas qualiEying higher acceptancec shorter e x p ~ t i o ndating period,re~geratort e ~ p e r astorage, t~ more protective c o n ~ n e r andlor closure,modifi~tionto the formulation and/ormanu€ac~ngprocess, should be considered during drug development. If CRT storage is ultimately justified, it may be necessary to add to the product labeling a cautionary statement against prolonged exposure at or above

300 301 302 303

30°C.

endix

304

307

The Iong-term testing will be continued forsuscient a period of time beyond 12 months to cover shelf life at appropriate test periods. The M e r a c c data ~ us h~ oa~ d be ~ ~ b m i ~ to e dthe FDA during the assessment period o f the drug application.@CH Q 1A] The firstthree productio~batches m ~ u f a c t ~ post e d approval, i f not submitted in the o ~ g ~ l a p p I i ~ t i oshould ~ be placed on accelerated and long-term stability studies using the same stability protocolas in the approved drug ~ppli~tion. [ICH Q1A] A m ~ u omf 4 test stations

and dry powder inhalers (DPIs). Further ~ f o ~ a t i about o n these products andc o n ~ n e ris s provided in this section. a. Stability Storage Conditions for Drug Products in Semi-Pe~eableand Permeable

3 ~ 0

321

324

Containers

For large volumeparenterals (LWs), smali volume parenterals (SVPs), o p h ~ a ~ cotics, s , and d sprays packaged ins e m i - p e ~ ~ bcontainers, le suchas plastic bags, semi-rigid plastic containers, ampules, vials and bottles with or without d~ppe~applicators, which may n d storage ~ w n ~ t i o nare: s susceptible to water loss, the r ~ ~ e stability Q

327 ~~8

330 331

* *

Accelemd condition: 40°C f 2"C/15% RH & 5% ~

W @ C HQ

eto as r4~"CI15%

For liquids in glass bottles, vials, or sealed glass ampules, which provide an i m p e ~ ~ b l e

b. Stability Storage Conditions for Drug Products Intended to be St0 37 ~ 3 9

Mer

Accelerated condition: 4O"C/mbienthumidi~is an acceptable altemative to40°C/75% RH; Inte~ediatecondition: 30°C/ambient humidity is an acceptable ~ t e ~ t i to v 3OoC/60% e M; Long-term condition: 2 5 " ~ ~ b i e humidity nt is an acceptabie alternative to 25'~60% M. Refrigerator T

Accelerated conditio^^ 25"C/60% mbient ~ ~ i d iant ya ~ e p ~alternative b ~ e for aqueous products that would not be by humidi~conditions; Long-termconditions: 5°C f 3"C, with m o ~ t o ~ nbut g ,notcontrol of, humidity. IO

referredto as 30°C/4~%

~ n t e ~ e d i acondition: te 30°C f 2"C/40% RH f 5% Long-term condition: 25"C f 2"C/40% RH rt: 5%

to water foss, 0

c. Stability Storage Conditions for Drug Products Intended to be Stored at Freezer Temperature

340 34I 342 343

A ~ l ~ t conditions: e d 5"C f 3 ' ~ ~ b i e n t long-^^ conditions: -15°C f 5°C.

h ~ d i ~ ;

. S ~ b i lStorage i~ Conditions for Some I ~ a l a t i o nProducts 345 346 347 8 349

7. TestingFrequency

350 351 352 353 354 355 356

Frequency of testing should besufkient to establish the stability characteristicsofthe drug p~duct.Testing will n o ~ a l l ybe every 3 months over thefirst year, every 6 months over the second year, and then mually. ~ a or bracketing ~ can ~ be used, g if just@ed. DCH Q1 A] A ~ ~ of 4 testr stationsn(e.g., 0,2,4, ~ and 6 m o n ~are ) r ~ o ~ e n d for e dthe 6-month accelerated s~~~~ study.

357 358 359 360 361 363 364 365

though the ICH Guidance for Stubili~Testing of New Drug Substan~esancl ~ r o ~ u capplies ts only to new molec~arentities and associated drug products, applicants may wish to voluntarily switch to the I C H - r e ~ ~ e n d storage ed conditionsas defmed in ICHQ1A and Sections II.A.4. ~ as described in Section II.B.6. and II.B.5. o f this guidance or otherF D A - r e c o ~ e n dconditions of this guidance, as appropriate, for previously approved drug or biologic prod^^. Applicants or batches intendedto are not required to make such a switch for either annual stability batches support supple men^ changes. Although the following discussions referonly to the ICH conditions, the samer ~ ~ e n d a t i ocan n s be applied whena switch to otherF D A - ~ o ~ e n d ~ d conditions is ~ntemplated.

366 367 368 369

Two plans are presented to assist a p p l i ~ t who s desire to switch their approved drug products to the I C ~ - ~ ~ ~ e storage n d e conditions. d Under each pian,reco~endationswill be made on how to initiate a switch to the ICH storage testing conditions, select batches, collect data, report results, and proceedif products fail the approveds ~ i ~ ~ t iunder o n stheICH conditions.

362

370 371 372 373 374 375

~ p p l i ~ t i oofnE H Stability Study Storage Conditions to Approved Appli~tions

e s t i n ~ ~ o N d i tfor i o ~8sx1Approved S t a b i l i ~

This pian may be most suitable for dmg products that have been c;onfirmedto be stable when exposed to the controlled levelof h ~ i doni a~long-term basis. Only one set of conditions (i.e., as defmed the XCH conditions) and one setof testing for eachof the three verification batches, below, are necessary underthis plan.

376 377 378 379

Drug Products with an Approved Stability Protocol

i ~ under ana~pro ~ p pwho~ have i previously ~ ~ ~~0~~ drug product s ~ b i lstudies ~rotocolat 25"C, 30°C, or 25-30°C without h ~ i d controls i ~ may switch over to the IC i o n g - conditions, ~~ &F defmed in VB.o f the ICH Q1A guidance and i n c o ~ oin~Section t ~

382

r 390 391

ii. Products ~ i t h o uan t A p p ~ o Sv ~

~93 394 395 396 97 98 399

batch^, may be studied

K. Stability Data for ~ u p ~ l e r n C en~ ~ e s

403 405

407 8 9

the ICH Q1A guidance (~ectionI1.B. of this g

~or Section ~ 1II.B o ~f

)

iv.Other Conside~tions

For a mois~re-sensitiveproduct, the &pplicant may wish to explore the~ s s i b iof ~ ~i ~ swi~h-over to the ICH CO the con~iner/closurebefore e m ~ on~the n ~

is an a c c e ~ ~ baltema~ve le to 25 ~ 6 0 % for ~ o n g - ~ ~ nt these conditio^ should not be usedas the basis for a labelings ~ t e ~ esuch or "Store at 15-30°C'' to gain marketingadvanta~e. With respect to on~oings

~ ~studies, i ~a p i p~ ~may i ~carry ~ them to completion under the 12

s ~ d i ~

414 415 416

p ~ v i o ~~pproved iy ~ n d i t ior o may, ~ for practical or economic reasons, choose to make i ~ ~ d i aswitch t e toICH conditions and report the change in the nextannual report.

v. Data Subm~ssionto FDA

424

438 439

i Products with a n A ~ ~ Stability ~ o~~ r o t~o ~d l A p p ~may i ~ at^ ~ ity studies der the ~ ~ ~ - r e c o ~ e n d e d l otestnin^ g - ~nditions, te~ in addition tothe p r e v i o ~ ~approved y conditions at 25"C, 30"C,or 25-30°C without h ~ d i ~ 13

452 453 454 455

controls, for three consecutiveannual batches. Data from these annual batches under the ICH conditions shouldbe used to verify the current expiration dating period. However,i f the applicant wishes toverify the ICH conditions over a shorter time span, three production batches within one year or less may be selected, instead of three consecutive annual batches. ii. Products without an Approved Stability Protocol

456

460 461 462 463 464

A p p l i who ~ ~have p ~ ~ o u sperformed ly stability studies on a drug p r ~ u c t ' approved protocol should submit anapprop~atep r o t ~8s l ap ~ o r - a p p ~ S v~ protocol should con^ 25 O ~ ~ b i e n t h ~ d i t y cond~tionsand the ICH long-term conditions,as accelerated testing conditions. Upon approval of the protocol, a p p ~ c a nmay ~ tia ate cons~utiveannual batches at both 25' ~ humidity~ and 25O~60% ~ ~ Data firom these ~u~ batches under theICH c o n ~ t i can o ~ be used to ,veri the current, or to establiih a new, expiration dating period. ~~~~

iii. Other ons side rations

65 466

Sam&as in Plan A.

iv. Protocol Revisions 469

P ~ o d u ~with t s an approve^ s t a b i l i ~ protocol:

470 71 472 473

A p p l i who ~ ~ trave m approved stability protocolmay s u b ~the t a l t e ~ t stabi~ty e protocol in the annual report, reflecting thetemperat~eand h ~ d i t as y r ~ o ~ e byn the ~ ICH. d changes to ?he stability protocolgener~1yshould be submitted in a prior- appro^ supplement, d e s s the changes are to comply with the current co~pendi~.

474 475 476 77 478

Once adopted as an alternate protocol through anannual report, the ICH used, in parallel with the previously approved conditions, to gene~te s~bility data for ~ b s ~ u esupple n t men^ changes. ~ t e ~ t i v e lthe y , a p p ~may ~ treport the alte~ativeICH conditions in a supplemGnt, which requires stabilitydata, i f the supplementoccurs before the next scheduled annual report.

479

If the completestabilitydata gene~tedonthe

~~~

mual batchesundertheICH long-term conditions usingan approved alternate protocol(asdefined above) supportthe previously approved e ~ i r a t i o ndating period under the non-I conditions, the altemate can be adopted as the p~~ s ~ ~p r lo t i~ l~ ~ oanu ~g hu a l s ~ b i lprotocol i~

~ r o ~ u cwit~out ts an approved s t a b i l j ~ protocol: 484 85 486 487

For a p p l i ~ t i that o ~ do notconkin an approved stabili~protocol as defined above, a newor revised stability protocol may be submitted in a prior-app~oval supplement marked expe~ifed i dprimary ity review r e q ~ e ~ fThis e ~ .protocol should encompass25O ~ ~ b ~ e n t has~the long-term storage conditions andthe ICH long-term conditions, as the alternate,as welt as 14

488 489 490 49I

accelerated stability storage conditions,as defined by the ICH Guidance and above, and other reco~endationsdescribed in Ehis guidan~.Upon approval of the protocol, stabilitys ~ d i ~ may be i~tiatedon annual batches and batches intended to support supplementalc ~ e s . v. Stability Data for sup pie men^ C h ~ g e s

492 493 494 95 496 7 498 499 500 501

502 503 504 505 506 507 508 509 510.. 511 512

S a t i ~ a cdata t~~ If the complete stabiIity long-term conditions us expiration dating period under the non-ICH conditions, the data can be submittedin the annualreport and the current expiration dating period can be retained.

~ ~ a t i ~data a c t ~ ~ If the stability data under the ICH conditions fall outside the acceptance ciiteria while data fkom the parallel study under the previously approved conditions or 2S"Uambienth ~ d i t y , whichever applies, are satisfkctory during the previously approved expiration dating period, and the added h ~ d i t is y determined to be the cause for the stability f ~ l u r ethe , product will still be considered to be in compliance with the regulatory specificationsapproved in the application. I f the applicant decides to adopt the ICH conditions, ~ a ~ g e s - ~ ~ ~ ~ f f e c t e d supplement with shortened expiration dating period orprior-approv~ a s u p ~ ~ m ewith nt revised product specifications may be submitted wherejustified. Other measures (e.g., more protective con~iner/clos~e or productr e f o ~ u ~ t i o may n ) be considered througha prior-ap~~ supple v ~ men^

513 514 515 516

A l t e ~ t i v e l y , a hcareful r considerationof all aspects, the applicant may decide not to pursue the s ~ t c h ~ vtoe the r ICH conditions for the product.The applicant may ekninate the dtemate stability protocol in the nextannual report if a full explanation, including all related data and investigational results,is provided.

517 518 519 520 52 l

In the case where the product fails to meet the specifications under the non-ICHcon~tions, irrespective of whether it also fails under the ICH conditions, a thorough investigation in accordance with C G should ~ be~ performed and appropriatecomctive actions shouIdbe taken, including a Field-Alert Reportand recallo f the &ected product from the market place if warranted-

522

9. Packaging Materials

@CH Q1A]

523 524 525 5 ~ 6

The testing shouldbe carried out in the final packaging proposedfor m~keting. Additio~~ testing o f the protected drug product can form a usefulpart o f the stress testing an eval~tion,as can studies carried out in other relatedp a c ~ g i n gmate~alsin ~ p p o ~ i n definitive pack(s). 10,

8 9 0 1

Evaluation PC€€QlA] s ~ o ~be l dopt^ in the presen~tionof the ev~uationof the stab i uld COveT, ~ c l u ~ g

rate for oral solid dose f o ~ ) .

536

If ~ m b i ~ data n g h m several batchesis ~ p p r o p ~ a tthe e , overall retest period mayde^^^ on the time a batch maybe expected to remainwithin acceptable andj~~~ lii&. ~~~

The name ofthe degradation cubic fhnction of an arithmeti to the test the goodnessof fit of the data onall batches andco~binedbatches (where approp~a~) assumed degradationline or curve. 50 51

Where the data showso little degradation and so Iittle v ~ a bthat ~ iti is~a p ~ ~ n t looking h m data that the requested shelf life will be ~ ~go to ~ s ~ the at the f o d statistical analysis; buta j ~ t i f i ~ t i for o nthe omissions h o ~ d be pm6

553 54 55 556 7 8

Limited e ~ p o ~ a t i may o n be taken of the real-time data beyond the observed range to extend p ~ c ~ where ~ l the ya c ~ l e ~ t ~ expiration dating at approval time, this assumes that the same data, and hence theuse of known about such factors as the mechanismof d e g ~ d a t i othe ~ ~oodnesso f fit of any mathematical model, batchsize, and existence ofsupportive data.

559

Any evaluation should cover not only the assay, but also the levelsof degradation products and 16

g s

appropriate a~ributes.Where approp~ate, a~ention should be paid to reviewing the adequacyof the mass balance, di~erents ~ b i l iand ~ ,d ~ r a ~ t i o n p e ~ o ~ a n ~ . product aRer reconstituti~gor diluting according to labeling shotildbe riate and s u p ~ ~ i v e i ~ o ~ a t i o n .

VIII.~.for a d d ~ t i o ~ 566

572

578 579 58Q

11.

i ~ o on ~ adrug t i o products n which are r ~ o ~ t ior~ t ~ d

Sta~rnen~abe~~

c ~ c t e r i s t i c of s the drug product.

a. Room Temperature Storage

S~tements

581 582 583

( 7 7 O ~ ) e; x c u ~ s i o~~ se r ~ toi l5~ e ~ [see USP Con~olledroo^ tempera^^]

587 588

For sterile waterfor injection (WIand ) LVP solutio^ of ~

osalts p ~

~ 9 Q s ~ m i - ~ ~ e acontainers ble (e.g., plastic bags) the f o l i o ~ nstatement ~ may be used onthe

591

immediate container labels:

592 593 594 595

Store at 25°C (77°F);excursio~sp e r ~ i t t e to 15-30°C (59-86°F) [see USP Controlled RoomT e r n ~ ~ ~ e ] (see insert for m e r i ~ o ~ a t i o n ) and the following statement may be used in the "HOW Supplie~'section o f the package insert:

596 597

599

600 601

605 606 607 60 609

610 611 612

613 614

615 616

617 618

619 620

621

LW solutiom packaged in a semi-permeable container (e.g., a plastic bag) and containing simple organic salts (e.g.,acetate, citrate, gluconate, and lactate, and dextrose10 percent orless) may be l a b e l e d as above, provided thereare adequate stability data'(& least 3 months' at 40°Cf 2"C/lS% M& 5% or 40"CMMT 20% REI) to support such labiling.

iL AI1 Other Dosage Forms For all &er dosage forms (e.g., solid oral dosage forms,dry powders, aqueous liquid, semi-soli4 and suspension dosage forms) that have been demonstrated to be stable at the ICH-recommended conditions (25°Ck 2"C/60% R €fIS%, or 3 O a ~ 6 O %RH & 5%) or at non-ICH conditions,such as 3O0C,25-30°C,or 25°C without humidity controls and intended to be stored at room temperature, the recommended labeling statementis: Store at 25°C (77°F);excursions ~ e r ~ i t t to e d15-30°C (59-86O~ [see USP Controtled RoomTempe~ture~

iii. Where Space on the Immediate Containeris Limited W e r e an abbreviated labeling statementis necessary because space on the immediate containeris l i i t e d , either of the following statementsis acceptable provided the.firf1 labelimg statement,as shown above, appears onthe outer carton and in the package insert: Store at 25°C (77°F);e x ~ u ~ i 15-30°C o ~ s (59-$6"F) Store at 25°C (77°F)(see i n s e ~ )

b. Refiigerator Storage Statement For a drug product demonstrated to be stable at 5°C f 3"C,2-SoC, or 2-8°C with or without 18

endix

622 623 624 625 626

h ~ d i t control y and which is intended to be stored at refrigerator temperature, the recommend^ storage statement for labeling may be oneo f the following:

StQrein a refrige~tQr, 2 4 ° C (36-46°F) Store r e f r i ~ ~ 24°C ~ t e ~(36-46"~ ,

629

l a ~ l i n gstatement is necessary because space statement is ~ p ~ l e shown above, appears on the outerc o n ~ e r

630

efriger~te(see inse~)

631 632 633 634 635 636 637 638 639 642

on the

c. Room Temperature and/or Re~geratorStorage ~tatement For a drug product demonstrated to be stable both 2 at5"Cf 2"C/60%RH *.5% and at refiigerator temperature, eithedor botho f the roomt e m p ~ r a and ~ e refiigerator labeling as described above, are acceptable, depending on the storage conditions intended for sta~mentsy the product. A statement such as "store at 2-25 'C" is not recommended. d.AdditionalCautionaryStatements

If warrantedy additional cautionary statements to protect a drug product from excessive heat,

light, ~ ~k z i n g , and i other~damaging~ conditions, , should be included on the wntainer label is too limitedto dqlay all the and the package insert. Ifthe space on the container label r e c o ~ e ~ statements ed in detail, a reference to the package insert M for e r idomtion (e.g., see insert) is recommended. The uniform storage statements and stabdity condiions are s e in Tables 2 and 3, respectively.

Table 2: S u r n ~ of a ~UniformStora~eStatements in Drug Product L ~ b ~ ~ i n ~ I

Intended st0

~ ~ i t i o ~ for drug

product

9 650 651 652 653 654

Table 3: ~onditionsunder w h i c ~ to Apply ~ n ~ o r m

een Shown to be Stable §to~~e §tat~~en~

Ref~ge~tor

Temperature

LVP in a plastic bag'

655

Ail drug products

(as~ p p r o p ~ ~ t e )

656 657 8 9:

660 661

5°C f 3OC 2-5"C ' Of 2-8°C

'Sce Section tIB.1 'Ia.for additionai informationon sterile water for injection and LVPs containing in0 or simple organic salts.

662 663 664 665 666 667 668 669 670 671 672 673 674

677 678 679 682

683

684 685

Other

Co~ideratio~

The a p p l i ~may t wish to include the de~nition of USP CRT in its entirety in the package insert to provide easy reference. f. Implementation of

USSs in Labeling New for Product

Applicatio~

30°C, 25-30°C, 25°C and at ambienth ~ i d i t y .

form is defmed as a drug product thatis a differentp ~ a c e u t i c aproduct l type, but c o n t a b the same active substance as included in an existing drug product approved by the FDA. New dosage form include products o f different a ~ t i o route n (e.g., oral, whenthe o r i g ~ uct was a p ~ n t e ~new ) , specific ~ c t i o ~ i t y / d e lsystem i v ~ (e.g., mod n the original new drug product was an immediate release tablet, md dosage forms of the same administration route (e.g., capsule to tablet, solution to ~ s p e ~ i o n ) . Stabiliv pro to^^ for newdosageformsshouldfollowthe guidance in the ICH However, a reduced stability database at submission time (e.g., 6 months'acceleratedand 6 months'lang-termdatafromongoingstudies)may be acceptable in certainjustified cases.

686 687 688 689 690 691 692

Stabiliv protocols for new combwtion productsornew f o ~ u l a t i (which o ~ require clinical data for approv~)should follow theguidanceinthe ICH Q1A in principle.However,are stabilitydatabaseatsubmissiontime(e.g., 6 months'acceleratedand 6 months' data from ongoing studies at the long-term ~ondition)may be acceptable in certainjustified cases, such as when there is a s i g n i ~ cbody ~ t of i~ormation on the stability of the drug product and the dosage form.

694

Much of the general i ~ o ~ a t i provided on in this guidance is app~cabletoabbreviatednewdrugs (AND&). However,dependinguponthe a v ~ l a b i l i ~ significant of information on, and the

695

complexity of, these drugproduct~dosageforms, the amountuof i ~ o ~ a t i necessary on to support these a p p ~ i c a t i may o ~ vary h m that proposed for NDAs. This section is intended to provide specific r e c o ~ e ~ d a ton i oabbreviated ~ app~i~tions.

696 697 698 99

71 711 712 713 71

appli~ble, interm~ate stability study storage conditions (referto 1V.G. of the ICH Q1A Guidance and Section1I.A. o f this guidance). Stability samples shouid be stored in the bulk of ~ n ~ n eclosure r , and liner, but storage ~ n ~equivalent e r (e.g., same composit~on and type smder in S&).

715 716 717 71 719

If not previouslygene~tedor available by reference,stress testing studies~ o u l be d conducted to establish thei ~ e r e nstability t characteristicsof the drug substance, the proposed procedures. The detailed natureof the stud ~ d i v i drug d ~ substance, type of drug product and available s u p ~ ~~f og~ a t i o Any n. ne cess^ testing m y be carried out as described in Section 1I.A. ~~~

720 721 722 723

727

Id contain stability datag~neratedunder the long-term and accelerated stability storage conditions delineated in V.E. of the ICH Q1A guidance (Section I1.B. of this guidance). The data package forANDAs (e.g., number o f batches, lengtho f studies needed at sub~is5ionand at approval, andacceie~ted,i n t e ~ ~ i aand t e long-term stability data) should be based on severaffators, includmg theco~plexity the of dosage€om, the ~ ~ ~o€an c e s i g n i ~body ~ t of in$ormation for the dosage form, and the existence o f an approvedappli~tion for ap ~ i c udosage l ~ form.

8 7 ~ 9

For Simpte Dusage Form the following stability data packageis r e c o ~ e n d e d

730 731 732

Accelerated stabili~data at 0,1,2, and 3 months. A tentative e x p i ~ t ~ odating n period of up to 24 months will be granted based on satisfactory accelerated stability data unless not supported by the available long-term stability data. 22

733 734 735 736 737 738 739

Long-term stability data (availabb data at the timeof original filing and subsequent

endm men^). A of one bate& pilot scale,

Ad~tionalstabi~itystudies (12 ~onthsat the int~rmediate conditions, or long-term data c ~ ~ is~seen g eafter3 months during the dating period wilt.be d e t e based ~ ~on

740

746 747 748 749

750 751 752 753 754 7~~ 756 757 758 759

An ANDA that is d e t e ~ ~ to e be d oneof the above categories should contain a modified ICH .. QlA stability data package,i n c l u d ~ : Q)

3-month accelerated stability studies. Long-term ~ i lstudies i ~( a ~ l a b l edata at the timeof original filing and subsequent en^). The expi~tion dating periodfor complex dosageforms will be determined based on available long-term stability data submitted in the application. A mi^^ of three batches ~ ~ a in accordance c ~ with d the ICHQ1A batch size ~ c o m m e n d a t i o(refer ~ to V.B. o f the ICHQ1A guidance andSection 1I.B. of this guidance). Additio~lstability studies (12 months at the i n t e ~ ~ aconditions te or long-terrn ~ ~ l i t y testing through the proposed e ~ i ~ t i date) o n if s i ~change ~ is tseen &e? 3 months d ~ n the g a ~ ~ e ~ studies t ~ (the tentative s ~ expiration ~ i l dating i ~ priod will be d e t e based ~ on the available data &om the additional studies).

760 761 762 763 764 765 766 767 768

Fu1l-term stability testingof the primary stabiiity batch(es) is suggested. However, in the absence of ~ l ~ - t e rstability m data for thedrug product, adequate accelerated stability data combined with ~ x p i ~ t i odating n period. available long-^^ data can be used as the basis for granting a tentative The batch(es) used for stability testing should comply fully ,with theproposed s ~ i ~ ~ for t i o ~ the product and be packaged in the market p a c ~ g eand , the releaseassay should be within reasonable v ~ a t i o n(taking into accounti ~ e r e nassay t v ~ a b i l i &om ~ ) the labeled strength or theoretical strength ofthe reference listeddrug. If formulat~with an overage, the overage should be justified as necessary to match thato f the reference listed drug*

769 770

Other supportive stability data may be submittedon drug product batches that may or may not meet the above criteria. Data on relevant research batches, investigational formulations, alternate 23

771 772 773 774 775 777 779

c o n ~ n e r / c l osystems, s~ or fiom otherrelatedstudiesmayalso be submittedtosupportthe ~ b i ~ iof t ythe drug product.Thesupportive s ~ b i l idata ~ should be cleariyidentified.

.Stabiii~Study Accepta~ce If the mults m satisfactory, tentative a expirati to

24 m o nthe at labeled ~ es me Used to projat a ies ~ ~ t i v e e ~dating i ~ tperiod i o n thatis beyond adate s u p ~ byo a~c t d l o n g - t e ~ s ~ don n batches, thea p p l i ~ i o nshould include a~ ~ ~to conduct e nlong-^^ t stability S until the tentative ~ on dating is determined. ~ ~ e ~ofi o n eriod is verified, or the a p p r o p ~ ation generated on at least three n batches tested according to the a p p r o v ~ p r outlined o t ~ l in the s ~ b i ~ t y VI. o f this guidance. ng o f the data should follow Section ~~~

783 784 785 786

wi prior to publication of this guidance should not n o d l y have include to ility data in cor&ormanw withtheguidanceupon ~ b ~ s s i ifothe n original appIi~tionwas with^^ due to non-stability related issues. H o ~ vi f~new , 'ability -diesareconductedto s u p ~ rthe t submission,suchstudiesshouldbeconducted as recommended in theguidance.

788 789 790

Much o f the following Sormation is talcen from the guidance for industry, Content and Formai of Investigatio~iNew Drag Appiicatio~~~~s~ for Phase I Studies of Drugs, I ~ c i u ~ i ~ g ~ e l l ~ h a r a c t e rTherapeutic ~e~ Biotechnolo~~erived Products (November 1995).

791 792 793 794 795 796 797 798 799

The regulation at312.23(a)(7)e m p W i s the gradedmQmo f ~ ~ a and controls c ~ g ~ o ~ a t Although i o ~ in each phaseo f the ~ v ~ sufficient g ~~ o r m o a ~~ should o n be S to ensure theproper i d e n t i ~ ~ t i oquality, n , purity,and~n~ of the ~ v e ~ g a t i o ~ drug, the amounto f information needed to achieve that assurance will vary with the phaseof the ~ v e s t i g a t i othe ~ proposedd ~ t i o of n the inve~gation* the dosage form, and the amountof i~ormationotherwise available. Therefore,~ t h o u ~ h sdata ~ b are i l irequired ~ in all phases of the INXI to d e m o ~ ~that t e the new drug substance and drug product within are acceptable chemical and physical limits for the planned duration o f the proposed c l i i d i n v ~ g a t i if o~ very ~ g data can be c o ~ s ~ n ~very g l limited. y ~ 0 r t - tests t ~ are proposed, thes u p ~ stability modifications to the methodo f p ~ p ~ t i of o the n newdrug substance and

03 04 805 06

~ o ~ that wilo n identification of a safety concern or i ~ u ~ c i edata n t to make an eval~tionof d e t y are the only m o n s for placing a trial on clinical holdbasedonthe CMC sectior~

A. Phase 1 24

810 11

I ~ o ~ a t i to o nsupport the stabilityo f the drug substance during thetoxicologic studies and the proposed clinical ~ d y ~ eshould s ) include thefollo~ng:a brief d~criptiono f the stability study and the test m e t h ~ used s to monitor the stabilityof the drug substance andp r e ~tabular i ~ ~ data based on representative material. Neither detailed stability data nor the stability protocol to be submitted,

12 13 814 81 816 817

r e p r e ~ n ~ t i vrial. e Neither detailed stability data nor the stabili~protocol need tobe submitted.

807 808

809

818 819

Whm significant d ~ m p o s i ~ during on storage cannotbe prevented, thecliical tGal drug product shouldbe retested priorto the tia at ion o f the trial and ~ o ~ a t i should on s u b ~ to~ show d that it will remain stable during thecome o f the trial. data available when thetrial sta13~.Impurities that increase by reference toprior human or animal data.

823 24 825

~evelopment drug of product f o ~ u l a t i during o ~ phase2 shouldbe based inpart on the a c c ~ stability ~ a ~ o~ ~ ~a t i gained o n &om studieso f the drug substance and its f o ~ u l a t i o ~ .

826 827 828

The objectives o f stability testing during phases 1 and 2 are to evaluate the stabilityo f the i n v ~ ~ g a t i o ~ f o ~used u l aint ithe o ~initial clinical trials, to obtain the a ~ t i ~ oo ~~ t i o n needed to develop afinalfo~ulation,and to select the most appropriate ~ n ~ande closure r (e.g., compatibility studieso f potential interactiveeEwts bemeen the drug s u b s ~ ~and s ) other o n be s~~~ and componen~ofthe system). This i ~ o ~ t i should during phase 2. Stability studies on these fo~ulationsshould be well underway by the end o f Phase 2. At this point the stability protocol for studyo f both the drug substance and drug be product should be defined,so that stability data generated during phase 3 studies will a p p r o p for ~ ~submissionin the drug application.

829

30 31 832 833 834 835

843 844

In stability testing during phase IND 3 studies, the emphasis shouldbe on testingf d f o ~ u l a t i in o ~their proposed market, packaging andm a n u f a c ~ gsite based on the ~ c o ~ e n ~andt objectives i o ~ of this guidance. It is recommended that thefinal s ~ b i l i ~ protocol be well defined prior to the initiation o f phase 3 IND co~iderationshould be given to establish appropriate V i e batches o f the drug substance and drug product and those o f the primary stability batches in support o f the proposed expi~tiondating period. Factors to be considered may include, for example, source, quality and purityo f various components o f the drug product,m a n u f a c ~ n g process o f and facility for the drug substance and the drug product, and use o f same containe~ 25

raft Not for ~ ~ l e ~ n t a t ~ o n 845

and closures.

~ ~ l iprotocol. t y

ality, and p ~ t of y a drug product are ~

e d a c ~ g ~ ~as well kas for p hey s i c i ~~or p ~ ~ o t i o ~

,stability studies should includethe drug product p

for ~ samples. The stability protocol may also include containers tosupport short-^^ storage prior to

863

865

The stability protocol should include ~ e t h o d o l ofor~ each parameter assessed during the e the drug product. The protocol should also address stability evaluationof the drug s u b s ~ cand analyses and approaches for the evaluation of results andthe determinationofthe e ~ ~ t i o n dating period, or retest period. Thes~bility-indi~ting me tho do lo^ should be validated by the ~ ~ a cand~described e r in sufficient detail to permit validation andor v e ~ ~ by~FDA o n la~~to~es. The stability protocol for both the drug substance and the drug product shouldbe designed in a manner to allow storage under specifically defined conditions, For the drug produe6 the protocol should support alabeliig storage statement atCRT, reegerator temperature, or$ieezer temperature. See Sections 11.3.5and 6. A properly designed stability protocol should include the following ~ o

T ~ ~grade c and a ~~ ~ a cof drug ~ esubstance r and excipients Type, sbz, and number o f batches Type, size, and source o f containers and closures Test parameters Test methods Acceptance criteria Test time points Test storage conditions Container storage orientations Sampling plan Statistical analysis approaches ande v a l ~ t i o ~ Data presentation Retest or expiration dating period (proposed or approved) 26

884 885 886

892 893 894 895

Stability commitment The useo f a l t e ~ t idesigns, ve suchas bracketing and matrixing, may be appropriate (seeS~tions VI1.G. and H.). At the timeo f a drug applicationapprov~,the a p p l i ~has t probably not yet m a n ~ the ~ ~ d subject drug product repeatedly on a production scale or accrued full l o n g - t e ~data. The e ~ ~ t i dating o n period granted in the original a p p l i ~ o nis based data, statistical analysis o f availablelong-termandother supp acceptable accelerated data foran AMIA. It is often derived from p

reports.

p ~ d u ~ batches o n re~rted ~u~ annual The stabdity protocol approved in the on app~cationis then crucial for the w ~ ~ a t i purpose.

896 897 898 899 900 901

A stabilityc o ~ i t m e nis t acceptable when thereare sufficient supporting data to predicta favorable outcome witha high degreeo f confidence, suchas when an application is approved with stability data available from pilot-plant batches, when a supplementis approved with that do not cover theh11 expiration dating period, oras a condition o f approval. This commitment constitutesan agreement to:

902 903 904

1. Conduct andor complete the necessary studies on the k t three production batches and annual batches thereafiero f each drug product, container, and closure according to the approved stability protocol through the expiration dating period.

905 906

2. SuWtt stability study results at the timeintends and in the format specified by the FDA,

907 908 909 910 911 912

3. W i t h ~ w from the market any batches foundto fall outside the approved specifications for

913

For postapproval changes, items2 and 3 remain the same and item1 becomes:

914 91s 916 917 918

1. Conduct and/orcomplete the necessary studies on the appropriate number of batches. The

919

The approved stability protocol should be revised as necessary to reflect updates to USP

~ c l u d the ~ g annual batches.

is a single occurrence that the drug product. If the appficant has evidence that the deviation does not affect the safety and efficacyofthe drug product, the applicant should immediately discuss it with theappropriat~ chemistry team and providejustification for the wntinued in the distribut~drug orb i o l o g i ~ d ~ ~ b u t i o nthat o fbatch.The change or d~terioration 21 CFR 314.81~)(1)~i) or 21CFR 601.14,respectively. product must be reported under

amount o f stability data supplied will depend on the nature ofthe change being made. A p p l i may ~ ~ determine the appropriate data package by consulting the Pos~pproval Changes sectionof this guidance (SectionIX.)and in consultation with the appropriate chemistry review team.

920 923 924 5 926

m o n o ~ or p the ~ current state-of-the-art regarding the typeo f p a ~ e t me o~n ~ t o ~ d , acceptance critexia o fparameters, such testthe and rnethodolo~ CS. However, other m ~ i f i ~ tare i od ~i ~ ~ g unti1 e dthe e x p ~ t i o time o f approval has been confirmed by long-term data from production batches. Once a sufilcient database is established from several p ~ u c t i o nbatches to c o the o~ginally ~ ~ e approved ~ ~ tdating i period, o ~it may be a p p ~ p ~ ato t em ~ theis ~ b~~ i t y p See ~t~ol. Section 1X.J.

ene~t 929 930 93l 932

933 934 35 936 937

applicatio~includmg the mualbatch(es), and to support ~ ~ pchanges. ~ The r data o should ~ ~ be presented in an organized, comprehensive, and cumulative format.

938 939 940 941 942

946

foIlo~ng i~ormation and datato facilitate It is suggested that stability reports include the decisions conwming drug product stability: 1 GeneralProduct Informatio~

Name, source, ~ u ~ sites,cand date ~ o f ~g ~ ~of d~~ a substance c ~ ande&g or biological product. Dosage form and strength, including formulation. (Theappii~tionshould provide a tableof o n been studied, the specific formulatio~under study. W e n more than one f o ~ ~ a t i has f o ~ ~ ~ o isna ~ p ~ b l e . ) o f and closure. SWers, ~ r n ~ s type, i t i s ~o w ~ , size, and adequate d ~ ~ p t i o ncontainer S&, and desiccants should alsobe identi~ed. 2. S ~ c ~ ~andt Test i o~ e~t h o d o i o ~

952 953 954

~~o~ation

Physical, chemical, and ~ c ~ b i o l o gattributes i c ~ andr e g ~ a t so ~ p ~ i (or~ ~ ~ references to NDA, BLA,PLA, or USP). Test me tho do lo^ used(orspecificreferenceto IND, A,NDA, BLA, PLA prior sub~issions,or USP) for each sample tested. I ~ o ~ a t i on o naccuracy, precision, and suitability of the m e ~ ~ o l o g (cited y by reference to 28

955 956 957

appropriate sections). M e r e appl~cable,a description ofthe ~ t e n c test(s) y for measuring b ~ o l o gactiviq, i~ inchding ~ i f i ~ tfori potency o ~ det~ination.

3. Study Design and Study ~ n ~ t i o ~

958 959 960 961 ~62

Desc~ptionofthe ~p~~ plan, ~ c l u d i ~ 0 Batches and n ~ ~ r 4 Container and closures and n ~ selec~ted. r Number of dosage wits and ~~e~ tests were c o n d u on ~ ~i~~ ts or on composites of i n d i ~ units. d ~ 0 Smpliig timepoints. 0 Testing ofdrug or b i o l o ~ ~ for the at on time of on directed on thelakliig) as well as their recommended use periods.. Expected duration of the study. Conditions of storage of the p ~ u cunder t study (e.g., temperam, h ~ d il i g~h ~ ,container orientation).

~~~

967

969

4. Stability D a ~ ~ o ~ t i o n

970 971 972 973 974 975 976 977 978 979 980

~ ~date. a c ~ Batch number (research, pilot, p~duction)and associatedm For antibiotic drug products, the ageo f the bulk active drug substanc%(s) ~ ~ athe batch. c ~ g Analytical da@ source o f each datapoint, and dateof analysis (e.g., batch,c o ~ ~ ~ composite, etc). Pooled estimates maybe submitted if individual data points are provided. Individual dataas well as mean and standarddeviation shouldbe wported. Tabufated data by storage condition. SUIo f information on previous formulations during product development. This s m a r y may be referenced (if previously submitted) and should include other containers and ciosures investigated.

e 0

e 4

5. Data Analysis

981 982

"he following data analysisof q ~ t i t a t i v eparameters should be provided

3 4 985 986

* 0

e 0

Evaluation o f data, plots, and/orgraphics. Doc~entationof appropriate statistical methock and formulas tised. Results of statistical analysis and estimated expiration dating period. Results of statistical tests used in arriving at ~ c ~ b i o l o g i cpotency a1 estimates. 6. Conclusions

9 8 ~ 988 989 990

~~~

e 9

Proposed expiration dating period and its j~ifi~tion. ReguIato~specifications ( e s t a b l i s ~of ~ tacceptable minimum potency at the time of initial reiease forfull expiration dating period tobe m a t e d ) . C, F ~ r ~ a t t i n Stability g Reports 29

992 993

Submitted i~ormationshould be cumulative and in tabular form. Exampies are provided on the following l i t and in Table 4.

94 995 996

Study Number

Container ~ ~ ~ s i t i o ~ s u p p l i ~ Closure ~ ~ ~ s i t i o ~ s u p p l i e r S~Suppli~ ~ f ~ S i ~ ~

1000 1001 1002

Length of Study

1003 ~0~

*Batcbcs Used in Clinical Studies and Biostudies(Specify) **Batches of Different Formulation

Specs Failums ReportingPeriod

Appli~tion

1005

Data

Stability

ode1

Pr~sentation

1006 1 ~ 7 Product Name Study N ~ b e r 1008 ~ o ~ u ~ o n 1009 1OlQ Dosage and S~ength Drug Product BatchN ~ b e r / ~Numbe~b n ~ l X01 1 Batch Typeand S i m 1012 ~ ~ ~ ~ ~ r / S i t ~ a t e 1013 Drug Substance~ ~ ~ ~ ~ r / S Number i t ~ a t c h 1014 Container ~om~sitio~Supplier 1015 Closure Compositio~Supplier 1016 S~~uppiier 10117 ~ a ~ ~ g e r / S i t ~ a ~ 1018 1019 s ~ p Plan l ~ ~ SpecFkationsand Test Methods 1020 1021 Storage Conditions 1022 Length of Study 1023 R ~ p o Period ~ ~ g Location of Data inAppli~tion 1024 1025 S~ of Data 1026 Data fhalysis 1027 Conclusions 1028 a Batchesused in clinical studies and biostudics(SW@). 1029 Batchesof different formulations. ~~~~

Table 4: (cont.)

1030 103l

10~2 1033 103

1036

P ~ d u cNameBtrength t Batch Number Date ~ ~ Date ~ a c ~ g

Study N~~

Batch Size a

Purpose of Study Date Study Started ~n ~~ner/Si~uppii Seal Supp~er

1037 1038 1039

A ~ b u ~ sMethod

1 Specification I

Time ~ o n ~ )

1 ~ 0 1 ~ 1

1049 1050 105 1

1. Introduction

tion on 501(a)(2)~) of the FederalFood,Drug, andCosmetic Act states h t a drug shallbe in ~~0~~

with good ~

~ p d c e to assme ~

~ c t i forothe~storage of drug products under appropriate~ n d i ~ o n s tern^^^, of h ~ d i t ym , cilight so the identity, strength, quality, and purity of the drug prbductsare not ~ e c ~ d .

10~8 106~ 1070 1071

o n d i ~ ~ u(21~ CFR r s on governing state licensingof wholesale p r ~ c ~ p t i drug 205.50 (c)) states thatall presc~ptiondrugs shall be storedat appropriate ~ m p e and ~ under ~ s a p ~ ~ p r i~nditions ~te in accordance with r ~ u ~ m e n tifsmy, , in the labelingof such drugs, or with ~ q u ~ r n ine the n ~current editionof an official compendium, such as the U ~ P The ~ . e n establish^ ~ for a prescription drug, the re ulation also states that i f no storage r ~ ~ r e m are as defined in an official pen^^, to help ensure thatits identity, are not adversely affected (21CFR 205.50 (c)(l)).

1072 073

07~ 1077

1079 1080 1082

‘J.D.Haynes,”Worldwide Virtual Temperaturesfor Product Stability Testing”,J.

Phnrm. Sei.. Vol. 60, No. 6,

927(June 1971).

2. Ca~cu~ation

1083 1084

There are a variety of ways to approximate a MKT. The FDA ~ w ~ e n that, d sfor m ~ ~repaclagers, a cand warehouses, ~ ~ all~datapoints obtained be inserted d~ectly into the

1086 1087 1089

recorden, or a high-low ~ e ~ o ~ e t e r .

1090 1091

1094

-AH R

1095 T,= the mean kinetic ~ m ~ rina"IS ~ e AH== the heat of activation, 83.144 W*mole" R =the universal gas comtant, 8.3144 x lom3mole"*"^*^ Ttx=the high t e ~ p in eO K during ~ ~ the 1' week TIL=the low t e m in O~K during ~ the ~ l* week T&= the high t e m p e r a ~ ien O K during then* week T&==the low temperaturein O K durilg the n* week n =the total number of weeks (i.e. 52) T = absolute tempera~rein "K O K = "C(Celsius) + 273.2 "K= [("F ~ ~ e ~ -32).0.555] e i t ) -I- 273.2

~096

~097 1~98

1099 1100 1101 1l02 1103

l1 1105 1106

Note that 83.144 ~ o u l e ~is~anoaverage l value based uponmany on organic reactions. Smce = 10,OOOo~the above equationeanbe s ~ p l asi: ~ ~ ~~

1110

1111

-10,000 2n 34

1112

3. Appli~ion

1113 1114 1115 1116 1117 118 1119 1120 1121 1122 1123 1124 1 125 1126

dosage f o m it ~y

&I ~

t i.fthe p ~ u c t qe ~ has i t been y adve~ly

1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137

drug productin each typeof immediate c closure proposed for~ k e ~promotion, g , or bulk storage. The ~ s s i b i lQf i ~interac~on osure and the~ ~ n t iintroduction al o f e ~ c t a b linto es the drug produc~f o ~ ~ a t i o during ns storage shouldbe assessed duringc o n ~ n e r / c l o s ~ e q ~ ~ l ~ tstudies i o n using sensitive andq ~ t i t a t i v eprocedures. These studies are ended even i f the container andclosure meet c o m ~ n dsuitability i~ tests, such as those outliiedin the USP for plastic ~ n and elastomeric ~ eor plastic~ closures.A draft guidance is available on this topic entitled~ ~ ~ ~ i sofsDocu~e~tation ion in Drug Applicatio~ for Container Ctos~re S ~ s t ~es~e d f othe r ~ a c ~ g i of nH g ~ ~ Drugs a n and~iologics(June 1997). 1. Conta~erand Closure Size

1138 1139 1140

Stability data for a given stren may be bracketed by o b ~ data g for the smallest and the largest c o n ~ eand r closure to be commercially marketed, provided that thei n t e ~ ~ i a ~ l e and design (Sec~onW1.G.). ~ n t a i n eand r closureis of ~ m p ~ b~mposition

1141 1142

Physician m o t i o samples ~ that are in different closures con d or si fiom in the s ~ b i lstudi i~ themarke should e beincluded closure systems may be included in b ~ c k ore matrixing ~ ~ S For solid oral dosage forms packagedin large containers (i.e., d i ~ b u t i o nto the patient) full stability studies should be perfo institutions orc o n ~ cpackagers t isanticipat~.Samples for s~bilitytesting at different time points may be taken fiom the same container. S ~ b i Idata i ~ also may be necessary when the ~nisheddosage form is stored in interim bulk containers prior to filling into the marketed package. Ifthe dosage form is stored in bulk containers for over 30 days, real-time stability data under

l146 1147 1148 1149

1150 1151

specified storagec o n ~ should ~ o be ~ gene rat^ to d e ~ o ~ ~comparable ate s ~ b i Itoi ~the dosage form in the ~ k e t e package. d I n t e storage ~ of the dosage form in bulk c o n ~ n ~ ~ should generally notexceed six months. The ~mputationof the expiration datingperiod of the

1156 1157

product in

1170 1I71 1172 173 1174 1175 1l76 11 77

~~~

contact

the upright o ~ ~ ~~ bt ii lstudies oi ~ m y be c o n ~ u e donly in the most&redid o ~ e n ~ t i o n , which is generally the invertedor on-the-side position. 3. Extractab1es and

A ~ o ~ t i o ~ A b s o ~oft Drug i oProduct n

Componen~

Specific extractables testing on B drugproduct is not r e c o ~ e n d Inverted ~. vepus upright stabili~testingduringpreapprovaland ~ ~ p r verification o v ~ is d l y adeqdate.Extensive testing €or e ~ c ~ b lshould e s be ~~0~~ as part of the q ~ ~ ~ tofithe o ~nn ~ e r / c l o s ~ e ~ ~ p o n ~ labels, t s adhesives, , colorants and ink (see previously cited ~ a c ~ g i ~n g d ~forc e

with various solvents, elevated d e ~ to be ~ a c~~ p d ~ b land e , that

Loss of the active drug substanceor critical excipientsofthe drug p ~ d u cby t inkmc

I 186 1 187 1188

~ ~ ~ i n ~ r / c ~ ~ ~ ~ ~ ~ n e n t s i~ part of the s ~ b i lprotocol. com~nents,as well as rnoni effectiveness). Excessive loss ofa c o ~ p o n eor ~ tchange in 8 p ~ ~ twill e result r in the f the drug producttomeetapplicable s p e c i ~ ~ t i o ~ ,

C. b~icrobioIog~cal ~ontroland Quality 36

~ of ~

1189

1. P ~ ~ ~ a t i v e s ~ ~ ~ t i v e n e s s

contain preservative systems to control baeria and wing m a n ~ a c Acceptance ~ . criteria shouldbe cations for the chemical content o f p ~ e ~ a t iatv the es life. ugh thep r o d ~shelf t

1190

1197

roach provides a margin o f d e t y ~~

the limit and 8. marginof error fbr theassays.

120~ 12 1207 1208 1209 1210 121l 1212 1213 1214 1215 1216

121~ 1218

For ~ ~ p p r testing, o v ~thef b tk e production batches should be tested with a microbid o f the challenge assay at the start and the endof the stability period and at one point in the middle s t a ~ ~period t y if the test period equals or exceeds two years. The first three pro~u~tion batches should be assayed for the c h e m i ~ ~ n of t ethe n tp ~ e at dl~a p p~r o p v~ a tpe~~ ~ Upon d e m o ~ o of n chemical content~ ~ ewith mimbial ~ t e ~e ~ t iinv ~ ~ s three productionbatches,chemical assays maybe to d e ~ o ~the~ t e of the specified c o n ~ n ~ tofi preservatives o~ for subsequent annualbatches placed into stability ~~~~

2. ~ i c r o b i o i o gLimits i~ for Nonsterile Drug Products

~ o ~ e rdrug i ~ products e that havespecified microbialh i t s for drug product releaseshwl testedfor c o n f o to~the ~ i i i t s at approp~ate,defmed time points studies. l’he USP ,provides ~ c r o b i o l o gtest i ~ methods for microbiall i i t s and c o n c ~ g ~ c ~ o b i attributes o i o g i o~f nonsterile drug products. 1223

1224 1225

3. ~

tAssurance ~ for~Sterile ~ Drug Products i ~

The stability studiesfor sterile drug products should include datafrom a sterility testof each batch at the ~ ~ i of the ~ test n period. g Additional testing is recommended to demons~ate 37

1226

maintenance of the integrity of the microbial barrier provided by the container and closure system. Thesetestsshould be performedannuallyandatexpiry.

1228 1229

~ n t e ~ t ytheomicrobial f barrier shouldbe assessed using anapp~priatelysensitive and ~ e q ~ t evalidated ly container and c I o s ~ integrity e established and documented to show the ~ oof ~ t a container and closure system. Then ~ of s ~~p l eto sr be tested should be s*mlarto the ~ p ~ ~ nq ~ge m eprovided nt in currentUSP “ S ~ r i lTests” i ~ . The samples that pass ~testingmay ~ forother t stability y testing for containerandclosure ~ point, butshould not be rem4 to s t ~ ~for ge ity sti in^. ~ o n ~ n e r i n t e ~ t ytests do not replace the currentUSP “‘sterility Tests” c 7 1 > or 2 1 product release.

1231

rogens and Bacterial~ n d o t o ~

l 1239 1 1 1242 1 1 1~45 1246 1247 1248

Drug products with specifiedh i t s for pyrogens or bacterialen dot ox^

time o f release and at appropriate intervals during the stability period.F products, testing at the beginning and the end ofthe stabilityt e s ~ ~ r i o d dosage forms containingdry materials (powder filled orl y o p h i l i i p r o d u c ~ ) solutions .~d packaged in sealed glass ampoules may need no ~ ~ initial tt time i point. o ~ Products containing liquids glass incontainers with flexible ~ n should be ~ tested no less than at the~ ~ and the~end of the n stability g testperiod. For test produres ~ l ~ Lysate and s p ~ i f i ~ t i o nrefer s , to the FDA ~ ~ j d e l i none ~ a l i ~ t i ofnthe ~ i m Amoebo~te Test as an End-Product Endotoxin Test for ~ u m a n ,andAnimal Parenteral Drugs,~iolog~cat Prod~cts,and ~ e d i c aDevices, l the USP “”Bacterial do to^ Test” , and the USP “Pyrogen Test” 4512.

1~49 0 1

3 4 1255

The design ofstability a study is intended to establish, ing a l ~ t e d ~ n u m b e ro f batches of a drug product,an e ~ i ~ t i dating o n period product ~ u f a under c similar ~ circumstax~ces. This approachassumesth from this smallgroup of tested.batchesextendto all fkture batches. shouldbe represe~~tive in all respects such 8s formulation, m a n ~ a c site, ~ gCO closure, m a n ~ a process, ~ g sourw and quality o f bulk material of the p o p ~ ~ oo n production batches and c o ~ with o all ~ quality S i ~ ~ of~the drug i op r~o d u ~

The design o f a stability study should take into#nside~tion the v ~ a b i l i ~~~~d~ o f dosage ~ t so,f ~ n ~~~ni anbatch, ~ md ~ of batc~esto ensure that the result in^ data for 1259 unit or c o n ~ e8r ~truly : represen~tiveo f the batch as a whole and to q~~~ the v ~ a b i l i ~ 1 ~ ~ &om 0 batch to batch. The degree of variability affkcts the confidence that a fiture batchwould 1261 remain within s p e c i ~ ~ tuntiI i o ~its exp~ationdate. 1. BatchSampling

1263 1264 1265 1266

Batches selected for stability studies should optimally constitute a random sample fiom the ~pulation production of batches. In practice, the batches tested toes~blishthe expi~tiondating

1267 1268 1269 1270 1271 1272 1273 1274 1275

2. ~ o n ~ eClosure, r , and Drug Product ampl ling

1276 1277 1278 1279

~ e l ~ ofo containers, n such as bottles, packages, and vials,fioh the batch chosen for inclusion in the s t a b i l i ~study should ensure that the samples represent the batch as EL whole. This can be accomplished by taking a~ d o sample m o f c o n ~ efiom r s the finished batch, by usinga stratification plan whereby at a random starting point every nth containeris taken &om thefill in^ or packaging line (n is chosen such that the sample is spread over the whole batch), or by some other plan designedto emure an unbiased selection.

1282 1283 1284

pre~ously Generally, samples tobe assayed ata given sampling time should be taken from unopened containers.For this reason, atleast as many containers shouldbe sampled as the n ~ b e r saxnpling of times in thes ~ i lstudy. i ~

1285 1286 1288

For products ~ c ~ g ine containers d intended ford i ~ n s i n by g a p h ~ to ~m~tiple y pa ti en^, or intended forr ~ a c ~ g i or n gp a c ~ g e din unit-of-use containers, samples previously opened containers. More than one container should be sampled study. The sampling protocol should besubmi~edin the drug application.

1289 1290 1291 1292 1293 1294 1295

Dosage units should be sampled fiom a given container randomly, with each dosage~unit v i n ~ anequalc~~ randomly, then because dosage units in otherparts o f the c o n ~ n e rdosage , units should be sampled fiom all parts o f the c o n ~ ~ eFor r . dosage units in this fashion,thelocationwithinthe c o n ~ efrom r which the samples weretakensho c ~ e n t e and d this ~ o ~ a t i included o n withthe test resul~.

1296 1297 1298 1299 1300

Unless theproduct is being tested for~ o ~ o g e n e i tcomposites y, may be assayed instead of individ~lunits.I f more than one containeris sampled at a given sampling time, an equal n ~ o f units from each container may becombiied into the composite. I f composites are used,their makeup should be described in the stability study report. The same type of composite should be used ~ o u ~ othe u stability t study. For example, if 20-tablet composites are tested initially, then 39

1301 1302

20-tablet composites shouldbe used ~ u g h o u t .If a larger sample at a given sampling h e is desired, replicated 20-tablet composites should be assayed ratherthan a single assayof a

1306 1307

e ~ ~ t i v e n e sthe s ~results , ~ ~ p ~ s o ~ .

13 12 13 l 13 1

tube ~p~ procedures. Where the largest

1315 1316

Semisolid drug productsin s b s that me intend^ for single use:need notbe tested for ho~o~e~ity.

1~17 131 1319 1320 ~321

mark&ed size is more than20 times the ho~ogeneitytesting of an inte~ediatesize is end^

~~~

3. S ~ l i Time ~ g

The samplet h e points shouldbe chosen so that my tion on can at a sufl5cientmuency to determine with reasonable assurance the n curve). Usually, the r e l a t i o ~ ~can p be adequately represented bya 1 on on an arithmetic or a l o g a r i ~ scale. c Sta~litytesting for long-term studies generally shouldbe ~

o at ~

1323 and yearly during the first year, six-month intervals during the second, 1324 products predicted to degrade more rapidly, for example, certain radiop 1325.. e sshould be shortened. Stability testing for acce16rated studies intervals between m p l i i g b ~ 3 ~ generally 6 shouldbe pedormed at a of four time points, including theinitial~ l i ~ 1327 he. ~~

1329

Freezing samples &er sampling for thec o n v e ~ e n ~~ o f ~ analysis~is not m uacceptable~ practice becaFe it m y cause delay in finding and respondor m y adversely affect the stabilityo f a product thatdoes not

1333

of the ~ d t ohf the den^ l i i t The degradation curveis estimated most precisdy, in terms VII.E.2.), around the averageof the s ~ p l i times ~g about the mean curve (Figure 1, Section included in the study. Tlwrefore, testingan increased n ~ b eof r repli~tesat the later~ p l ~ g times, p ~ c u l a r i ythe latest sampling time,is encouraged becausethis will increase the ~ ~ o d . sampling t h e toward the desired expiration dating 4. AnnualStabilityBatches

1337

AAer the expiration dating period has been verified with three production batches, a testing 40

1338 1339 1340 1341 1342 1343

p r o g r for ~ an app~veddrug product shouldbe implemented to confirm on-going stability. For every approved a p p l i ~ t i oat~least one batcho f every strengthin every approved

13 1345

medid gases, blaod, or blood

1346 7

1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360

a high degreeo f ~ ~ d e nan ce The methods describedin this section are used to establish with during which average drug product attributes such as assay and the batch will remain withinspecifications. This e ~ i ~ dating o n period r n ~ ~ a c process ~ n g for the should be appli~ble to all future batches produced by the same drug product.

1361 1362 1363

I f an applicant choosesa n expiration dating period to ensure that the c ~ c ~o f a~large c p r o ~ ~ of o the n ~ d i ~dosage d ~ units i are wid& s ~ i ~ ~ diffxent t i o ~s~~~ , methods than.those proposed below should be considered? In thii settingy testingof i n d i ~ unitsy d~ ~ m ~ s i t ~be, h p ~ m rtty ~nk

1365 1366 1~67

~ p p ~ ~ t s to usw e a istatistid s ~ g procedure otherthanthose discussedin this guidance should consult with the c h review~team prior ~ to the ~initiationo f the stability study and data analysis. 2. ~ x p i ~ t i Dating on Period for an I n d i v i d ~Batch

1368 1369

1370

The time d ~ n which g a batch may be expected to remain within ~ i ~ ~ t idepends o n s not only also on the initial average value on the rateo f physical, chemical or microbiological changes, but

1969.

RG. Easterling, J. Am. Stat.Assoc., "Discrimination Intervais for Percentiles in Regression"',64,1031-41, 41

1371 1372 1373

for the batch. Thus, information on the initial value for the batchis relevant to thed e ~ ~ i ~ t i o n of the allowable expimtion datingperiod and shouldbe included in the stability studyreport. P e ~ n ~ g elabeloclaim, f not percentageof ti^ avenge value, is the variableof interest.

1374 1375 1376

The exp*~tion dating period foran individual batch is based OR the observed pa^^ of the q ~ t i t a t i v attributes e (e.g., assay, degradation products) d e r study and the precision by which it is e ~ ~ .

1377 1378 1379 1380 1381 1382 1383 1384

An acceptable app~achfor analyzing an attribute thatis expected todeem with timeis to

1386 1387

analyzing m attribute thatis expected toi n c m e with tixne, the 95 percent one-sided upper

confidence limit for the mean is m m e n d e d .

When analyzing an attribute with bothan upper and a lower speiificationWI~, special wes may lea&to application of a two-sided95 percent confidence l i i t . For example, although chemical degradation of the activei n ~ ~ i einn at solution product would cause a decrease in the assayed Figure l: S t ~ t i sAnalysis t ~ ~ of Long-Tern tab^

111

Time (months)

1388 1389 1390 1391 1392

c o n c e n ~ t i evaporation o~ ofthe solventin the product (through the~ n ~ n e r / c l owould s~) result in an increase in the concentration. Because both possibilitiess h o ~ be d taken into account, two-sided confidence limits would be appropriate.If both mechanisms were acting, the concentration might decrease initially and then increase.In this case, thedegradation pattern would not be linear, and morec o m p ~ i statistical ~ t ~ approaches shouldbe considered.

1393

If the approach presented in this sectionis used, average parameters such as assay and 42

1394 1335 1396 1 1 1 1

degradation pducts of the dosage units in the batchcan be expected to remain within s ~ i ~ ~to the t iendoo f the ~ e ~ i r a t ~ odating n period at a confidence levelo f 95 percent. “he expiration dating period should notbe dete~inedusing the p i n t at which the fittedl ~ t - s q ~ s line i n ~ ~thet ~sp r o ~ ~ in^ a t eti on knit. This approach is BS likely to o v e the ~ ~ batch average the am to period 8s to ~ d e ~it, * t e squares l i e is used with c i f i at~expiration ~ ~ ifthe

1401 1402

1406 1407 1408 1409 X410 141 1 X412 1413 14 14 1415 X416 14 17 equality for 1418 1419 X420 1421 1422 1423 1424 1425 1426 1427 ~ 4 2

above,such as the ~ ~ p t i o n

~ p t i ~derlying o ~ th ility ofthe ~~~d~ units

remains ~

oover the ~

d evenwhen t h h ~app~ach t m y be e violation o f the ~ s ~ p tin itheo data ~ is noted, an ~ a ~ ~ pthe~ o bs j ~ht i y of e d~~ an e ~ ~ t i dating o n period with a high degree of confidence. 3. ~

a Dating ~ Period o n for All Batches

~ b i i that i t yis, the ~ l a t i o ~ hbetween is asmall, ip the paraineterof interest p r ~ u and c ~timeis essentially the sarne from batch to batch be ~ ~ b i m into e d one ovedl estimate. Combining thedata should be s t a ~ ~data li~ sho~d suppo~edby p~~~ testing of batch s ~ i ‘fiel simildty ~ ~ of~the estimatede w e s on^ the batches tested shouldbe assessed by ~ p l y statistical ~ g tests of the ~~i~ of slopes and of m time intercepts. The levelo significance ofthe tests, expressedin the p-value, should be chosen so that the decision to combine thedata is made only i f there is strong evidence in favor o f c o m b i ~ gA. pwlw o f 0.25 for p r e l i statistical ~ ~ tests hasbeen recommend^' I f the tests offorslopes and el of si^^^^ o f 0.25, the data b m p-dues less than 0.25, a judgme~tshould be made BS to ether pooling couldbe permitted. The a p ~ FBA ~ ~~i~ p ~ review ~ &am should be consulted regding this d e t e ~ t i o n .

I~~tch-t~~tch

I f the ~e~~ statistical test rejects theh y ~ t h e of s ~batch similarity because of unequal i n t e ~values, t it may still be possible to establish that the lines are p ~ l e(i.e., l that the slopes are qual). If sa, the datamay be combiied for the purpose of estimatingthe common slope. The vi^^ e x p - ~ ~dating o n periodforeachbatchinthestabilitystudy may then be bY c o ~ i d e ~the i n initial ~ values and the commonslope using approp~atestatistical m e ~ o d o ~ o ~ . data fkom severaI batches aremmbiied, 8s many batchesas fmible should be combmed because c o ~ d e n c Iits e about the estimated curve will become narrower 8s the number o f batches If it is ~a~propriate to combine ~increases, usuaIly resulting in a longer expiration dating period. ~~~

‘&K.. fin,T.Y.D. Lin,and RE. Kelley, “Stabilityof Drugs: Room Temperature Tats”, in Statistics in flze F h a ~ ~ u c e ~ Industry, t i c a ~ d.CR. Buncher andJ-Y. Tsay, pp.419-444, Marcel Dekker, Inc.: New Yo&, 1994,

7T.A. BmmR ‘‘Analysis and Inference for ~~omptetety Specified Modelslnvolvingthe Use of Preliminary

Te~t(s)ofsignificance,”Bionretrics, 20(3). 427-442, 1964.

1429

data &om several batches, theovedl expiration dating period will depend on the ~ i n i r time n~ a batch may be expected to remain within a c ~ p ~ b l e l i m i ~ . 4. P ~ u t i o nin s polat at ion Beybnd Actual Data

The

methods for d

~ m exp ~ g

1437

data become available.

1. ~ m p u ~ t i o nExpiration of Date

1450 1451 1452 I453 1454 1455

The ~ ~ p u ~ otf the i eo ~~ ~ dating ~ operiod n o f the drugprodu~tshould gin no later the time o f quality control release o f that batch, and the date o f ase should ~ e n e ~not ly e gene^^ ~ in a exceed 30 days h m the production date, regardless o f the p a c ~ g ~ g .~ sup^^ o f the assigned e x ~ i ~ t i dating o n period shouldbe fkom long-term studies under the storage conditionsr ~ o ~ e n d in e dthe labeling.I f the expiration date includes only a month and year, the produ.ct should meet specifications through the last day ofthe month.

1456

In general, properstatistidanalysis of long-term ~ b i l i t ydata col

1458

Section W1.E. and exemplified in Figurel , should supportat least period. Exceptions do exist, for example, with short h~f-liferadio

1460

I f the production batch contains reprocessed materi~,the expiration dating period should be computed from the dateo f m ~ u f a of c the ~ oldest reprocessed materid used. a. Extension o f Expiration Dating Period

An extension ofthe expiration dating periodbased on full long-term stability data obtained Eom

at least three production batches in accordance with a protocol appro~edin the appi~cation may be described in an annual report (21 CFR 314.70(d)(5). The expiration dating period may be 44

1465 1466

e ~ ~ n d in e dan ~

1467

~ t e ~ t i v e l y the , istability f studyon

l rtr onlye ifthe criteria set forth in the app~ved stabili~ protocol data, ~ c l u d statistical ~g adysis, if ~p~priate. at 1

p i l o t - ~ batches e is continuedafterthe

1470 1471 1472 1473

1479 1480 1481 l482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493

1495 1496 1498 1499 1500 l501 1502

supple men^

2. Retest Period for Drug from A retest period for a drugsubs^^ m y be established based on the available data long-term stability studies and,as such, can be longerthan 24 m o n i~f sup^^ by data. A retest date should be placed on the storage ~ n ~ande onr the shipping~ n ~for ea b r .A drug substance batchmay be used ~ t h o uret& t during an approved retest period. However, beyond the approved retest period, any remaining portion of the batch should be retested ~ e d i a t e l before y use. Retest o f different portions of the same batch for use at is acceptable, provided that the batch has been stored under the defined different timesas n ~ d e d #nditio~,the test methods are validated and stabili~-indi~ting, and all s ~ b i l i ~ - ~ l a t ~ attributes are tested andtest results are satisfacto~. Satisfacto~retest resultson a drug substance batch retest period can bee ~ e n d e dfor that batch or any other batch. The purpose o f retest is to qualifu a specific batch of a drug substance for usein the ~ a n u f ~ ocf a~drug e product, rather thanto 45

1503 1504 1S05

recertify the drug substance with a new retest date. To extend the retest period,full long-term data from a formal stability study on three production batches using a protocol approved in an appli~tionor found acceptable in aD m should be provided. Similar to the extensiono f an expiration dating period fora drug product, a retest period for a otiginal application. This cm be extend^ beyond whatwas ity data (i.e., c o Q e ~the g an mual report based on M desired retest period on three production batches using ana p p r o v ~stability protocol).

510

l511 1512

In a w e where testing reveals al i i t e d shelf-life for a drug substance,it may be i~pproptiateto W a retestdate. An expiration datingpero i d,rather thana retest period, should beestabl~hed ~, ~bstan~). for adrug substance with alimited shew-life (e.g., some a n ~ b i o t ibiological 3. Holding Tmes for Drug Product inter media^

513

1514 1515 1516 1517 1518 1519 1520 1521

the finished drug product~ u f a c ~ with e d suchan i n t e ~ e ~should ~ t e be m o ~ ~ r on ed long-term stability. When previous testing o f an inte~ediateor the related drug product batches suggests thatan intermediate may notbe stable for30 days, the h o l ~ time g should be kept to a atxiqualified by a p ~ r o ~ astabi~ity te testing.

1522 1523 1524 1525

The fkquacy of testing of an interme~ate Where practical, testing shouldbe done at a ofan i n t e ~ ~ i a t At e . a mi^^, all critical parameters shouldbe evaluated at releaseof an intermediate and immediately priorto its use in the r n ~ ~ aofcthe~ ~e ~ drug product, e d

1526 1527 1528 1529 1530 1531 1532

In the event that the holdmg time for an inte~ediatehas not qualified by ~~ropriate to the related f ~ e drug d product batch should stability evaluations, the expiration date assigned be computed from the quality control release date of the intemediate i f this date does not exceed 30 days from the date o f product~ono f the interm~ate.If the h o ~ ~ ~hasgb e a q ~ h i fei e d by app~priatestability studies, the expiration date assigned to the related ~ h e drug d product can be computed fiomits quality control release date i f this release date does not exceed 30 days firom the date that thei n t ~ e ~ aist introduced e into them a n ~ a cotf the ~ finished drug product.

1533 1534 1535 1536 1531 1538

1. General

The useo f reduced stability testing, suchas a bracketing design, may bea suitable alternative to a fill testing p r o g m where the drug is available in multiplesizes or strengths. This section discusses the typesof products and submissions to whicha b ~ c k e t i ndesign ~ isappli~bleand the types of factors that can be bracketed. Applicants are advised to consult with theFDA when 46

1539 1540 1541 1542 1543

1544

questions arise. 2. A ~ l i ~ b i l i t y

The factors thatmay be bracketed in as~bilitystudy are outlined in ICHQ1A and described in detail below. The types of drug products and the typesof submissions to which b ~ k e t i n gdesign can be applied we also discussed. a. Types o f drug product

1545 1546 1547 1545

~ r a c k design ~ g is appli~bleto most typeso f drug products, includmg~ e d i and a ~ solids, liquids, semi-solids, injectables. Certain typeso f dmg products, such as m e t ~ d ~ o s ~ e d e r (MDIs), s dry powder inhalers@PIS) and transdermal delivery systems (TDSs), may not be a m e ~ b l to, e or may need additional justification for, bracketing design.

1550 1551 1552 1553 1554 1555 1556 1557

%%eke a range o f ~ n ~ n e r /sizes f i l lfor a drug product o f the same strengthis to bee v a l ~ e d , b r a c k e ~ gdesign may be appli~ble the if material and compositionof the c o n ~ eand r the type of closure are the same ~ u g h o uthe t range. In a casewhere either the containersize or fill size varies whiIe the otherr e m a i n sthe same, bracketing design may be applicable i.sithout justification. if In acase where both container sizeand fill size vary, bracketing designis appli~ble app~priate ju~ficationis provided. §uch justification should d e m o ~ ~that t e the variousaspects (surface ) the arealvolume ratio,d e a d - s p a ~ / v o lratio, ~ e containerwall thickness, closureg e o m e ~of i n t e ~ ~ asizes t e will be adequately bracketed by the extreme sizes selected.

1558 1559 1560 1561 1562 1563 1564 1565

%he= a rangeo f dosage strengths fora drug product in the same~ntainer/clos~e” (with identical materialand size) is to be tested, bracketing designm y be applicable if the formulation is identical or very closely related in ~ m ~ n e n ~ ~ m ~ s Examples i t i o n .for the former include a tablet range madewith different ~mpressionweights of a common granulation, or a capsule range made by filling diEerent plugfill weights o f the samecom~sitioninto different size capsule shells. The phrase vefy cZosely relatedf o r ~ ~ l a rmeans i o ~ a rangeof strengths with asimilar, but ,basic ~ m ~ s i t i such o n that the ratioof active ingredient to excipientsremains relatively ~~t ~ o u ~ h othe u t range (e.g., additionor deIetion of a colorant orflavo~g).

1566 1567 1568 1569 1570 1571 1572 1573

In the case where the amountof active ingredient changes while the amount o f each excipient or the total weightof the dosage unit remains constant, bracketing may not be applicable udw justified. Such j~tificationmay include a d e m o ~ t i o n comparable of stability proffie among the di~erentstrengths based on data obtained from c ~ ~ ~ ~ d e v e l o p batches, m e n t p ~ a r stability y batches, and/or production batches in supporto f primary stability batches,~ ~ i ~batches, e n t and/or annual batches and batches for postapproval changes, respectively. With this approach, the fo~ulationsshould be identical or very closely related, and thecon~iner/clos~e system should be the same between the supportive batches and the batchesfor which the bracketing 47

X 574

design is intended.

1575 1576

If the f o ~ ~ a t i iso sn i g ~ ~ ~ t ~ y among d ~the~ de ~r~ e r~e ntt (e.g., tion on or de~etion excipient, except colorantor flavoring), bracketing is gene~Iiynot appli~ble.

1577 1 1

Due to the corn Iexi in product f o ~ u l a app ~o~ c~ernis~ advance when questio~ b ~ c k e design. ~g j ~ ~ c a t i is on

=vi

1582 roval c o ~ ~ ebatches, n t

app~c~on. 1~94

1597

1599

1603 16

1606 1607

A bracketing design thatis not # n ~ e din the approved protocol in thea p p l i ~ t i is o ~subject to supple men^ approval (21 CFR 3 14.70~)(2)(ix)) (601.12).

to support two different chemistry, the two proposed changes couldbe combined into onep r i o r - a p p ~ vsupplement ~ even though qualib fora c h a n g e s - ~ ~ g - e supplement or annual reportunder 3 14.70 (c) or (d) or 601.12, or relevant consult the ~propriateAgency review a c ~ p ~ b of i l the i ~ new b~cketingdesign prior to thetia at ion o f the stability studies, and subs~uentlysubmit the data to support thep r o p o s ~change ~ o u the g appropriate ~ filing m ~ c ~ ~ 3, Design

A b~cketing proto~l should always include the extremeso f the int~nded~ ~ e ~sizesi and/or a l P h y s i c i ~samples or bulk pharmacy packs intended be to repac excluded &om the b ~ c ~ e t i n protocol g for # ~ e r c i a lsizes, but could be studied under theirown e or more, of protocols, i f a p p l i ~ b ~Where e . a targen ~ b e rfor , e x ~ p i four 48

be subjected to the sme type of iple and p ~on ~ l ~ a b ~~ i ~ e hilarity before ~

m b em ~ g

exlremes are found to be d i ~ i ~ l ~ , more stable than the least stable

1637

extreme:

1638

.M a t r ~ i n g

1639

1. Gened

when it isap~opnate encouraged before the

l645

1650 l651 1652 1653 1654 1655 1656 1657 ~658 1659 1660 166I 1662 1663 16 1665

a. Types o f drug product ~a~ design is applicable to most typesof drug products, including immediate- and modifiedrelease oral solids, liquids, semisolids, injectables. Certain typesof drug products such as MDIs, DPXs, and TDSs may not be amenable to, or may need a d d i t i o ~ justification for, m a w design.

b. Factors Some o f the factors thatmbe xnatrixd include batches, strengthswith identical fornulation, container si=, fill sizes, and i n t e r m e ~ time a ~ points. With justification, additional factors that can be matrix& include strengths with closely related fornulatio~container and closure suppliers, container and closure systems,o~entationsof container during storage, drug substance ~ ~ u f a sites, c ~andg drug productm a n ~ ~ sites. c ~ For g example, to just@ matrixing across HDPE bottles and blister packs, a tablet dosage form could be shown not be sensitive to to e ~ ~ e n tand s )that it moisture, oxygen, or light ugh stressed studies, including open-dish is so stable that the protective natureo f the ~ n t a ~ e r / c l osystem s ~ e made littleor no diRerence in the product stability (through supportive data). Alternatively, it could bed e ~ o ~if ~ ~ d , For additional info&ation on bracketing studies, seeW.R Fairweather, T.-Y. D. Lin, and R Kelly, “Regulatory, Design,and Analysis Aspectsof Complex Stability Studies,”J .Phurm. Sei., 84, 1322-1 326, 1995. 50

1666 1667 l668 1669

a~ropriate,that thereis no diEerence in the protective nature between twothe distinctively is needed to ensure that the ma~~ng different ~ n ~ e r / c lsystems. o s ~ Thej~ti~cation two otherwise protocol would lead toa successful predictionof the expiration dating period when diffmnt c o n ~ e r / c lsystems o~ are studied together.

1670

points, a ~ b u t e( stest (i.e., diEerent excipientsor

1671 16 ~ 2 1673 1674 1675 1676 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 l689 1690 1692

c. Data ability and Product Stability

The a p p ~ ~ ~ofl i~t ya ~ g . d to e primary s i ~ stability batches depends on the product ~bility and datav~abilityd e m o ~ ~ tthrough e d clical or develop^^^ batches. Data v ~ ~ i l i t y refers to the v ~ a b i ~oft ysupportive stability data within a given h t o r (i.e., batch-~batc~ ~ e n ~ - t osize-to-size) - ~ g ~and across diEerent factors (e.g., batch vs strength, strength vs size). It isassumed that thereis very little variability in the methods of stability samples.~ a ~design n isgapplicable if these supportivedata indicate that the small ~ a b i l i t y .Where the product displays product exhibits excellent stability with very ~ o d e ~ t e ~ with b i l imoderate ty variabilityin the supportive data, ~~g design is applicable with ~ d i t i o ~ l j ~ Conversely, i ~ ~ t i o nif .supportive data suggest poor product stability with is not applicable. Similarly, whether or not matrixing design m large varhbility, xnatrixing design be applied to postapproval c o ~ ~ ebatches n t or supplemental changes will depend on the cumulative stability data on developmental batches, primary batches, and/or production batches, as appropriate. ~~~

Table 6 illustrates the range of situations under which matrixing design is applica~le,if applicable justified, qenerally not applicable, and not applicable. The table is intended, in a qualitative manner, to serve as a general guide for sponsors when d e t e ~if~matrixing g design is approp~te for a drug product with respect to the likelihood that such a design would in aresult successM prediction of the expiration dating period. It does seek not to q ~ t i ~ ~ vdefme e l ythe different degrees of product stability or data variability.

1693

Table S: A~~licabili~ of

1694 1695

1696 1697

1698 1 1 1701 1702

1703 17

d. Types of submission Same as Section V11.G. 1.c.

1705

3. Design

1706

General a.

1707 1708 1709 1710

as well as at least two additional time points through the first12 m points i n c l ~ the ~ ginitial and 12-month t h e points. This original application contains lessthan h 11 long-term data at the tixneof~ b ~ s s i o ~

1711 1712 1713 1714 1715 1716

~~~g d ~ edesigns ~ n t for Although ~ a t rshould i ~ not ~ beperfiormed acrossa ~ b u t ~ , diEerent aabutes may be suitable where different testing~ e q u e n ccan i ~ be justified. Likesse, each storage condition should be treated separately underits own ~~g design, if applicable. Care must be taken to emwe that there are at least three time points, includ~gi~~~ and end points, for each binat at ion of factors under an accelerated ~ n d i I~f ob r~a c k e ~isgj ~ t i ~ ~ , the matrixing design should be developed afterward.

1717 1718

NI samples should be placed onstabiii~i~cludingthose that are not tobetested under the matrixing design. Once the study begins, the protocol should be followed without dev~ation. The

For original applications, am a ~ design g should always include

1719 720 1721

only exception is that, i f n ~ e sit is ~ a , ~ p ~ btol revert e back toh11 stability testing duringthe

study. But once reverted9 thefull testing shouldbe carried out through expiry.

b. Size of mtrixing design

1722

1724

1725

1726 1727 1728 1729 1730 1731 1732 1733 1734 1735

1736 1737 1738 1739 1740 1741 1742

m b o f ~h t o r~and s o a substantial ~ amounto f supportive data are available, there are 3~3x3~ ~ n gcouldbe as small as one half of thatof a full testing protocol. the size o f the r n ~ ~ design ~ ~ , ~ u c f'/z~means u ~ that u Zonly one half o f the total numberof samples in the~ ~ s p o n d ~ g 1 1 1 protocol will be tested under ther n a ~ n design. g Refer to Examples2 and 3 below for two designs with an overall size o f 5/12and %, respectively.

Combinatio~of

ount of S u p ~ ~Data' v e Available

1754 1755 17~6 1757

Excluding timepoints.

175

No mutriring means that~ ~ x i isnnotgsuitable.

175~

product.

i n i ~ u d ~ ~ o pbatches, r n ~ t primary to fonn the basisto support the stab

c, Statistical C o ~ i d e ~ t i o ~

n e design should be well balanced. An e s ~ a tof e the p ~ b a bthat ~ istability ~ '~t~mes from the matrixed study would be the samefor a given factoror across different factors should be provid~ if available? 3

d. Examples

design with five-seventh't h e points (overall S& ~ ~ Example ~ #l. Complete n g sevenths of full testing protocol)

five-

'For additional information on matrixing studies seeW.R Fairweather,T.-Y. D. Lin and R. Kelly, "Regulatory, Design, and Analysis Aspectsof Complex StabilityStudies," . LPhurm. Sci., 84,1322-1326, 1335. 54

1766 1767 1768

The following example (Table 8) involves a complete design of 3x3~3combinations of factors with five-sevenW time pointsfor a capsule dosageform available in 3 strengths of a common g ~ u l a t i and o ~ packaged in 3 con~ner/cios~e systems andlor sizes: Cl,HDPE PE bottle, 100 c o u n s t;and bli~er-pack.A 24-month expiration mb~tio of~fktors will be e points; thus the overallsize of design is 5M ofthe corresponding full testing protocol.

1774

1775 1776 1777 1778

1779 1780

1782 1783

Matrixing Example#2. Two-thirdshctional design with five-eighths time points (ove size: ~ v e - ~of efilli testing ~ p ~ t ~ l ) The folio^ design(Tab1 ex.ampIe of 32~3x3 ~ ~ b i ~ oftfkctors i o with ~ dosagee form wbi& is a ~ i a ~inl 3estrengths of a c o ~ g o~ ~~a t i and o np a c ~ ing3 ~ c o n ~ ~ / c l o ~ m and/or s b . Cl, HDPE bottle; 30 ~~~; C2, HDPE bottle, 100 E bottle, 200 counts. A 3 ~ r n o exp~tion n~ dating period is pmpos size of this design can be r e f d to &S U3 (of 27 ~ ~ b ~ oftf~tors) i ox 5/8 ~ (of 8 time in^), or 5/12 (of 216 samples in 8 full testing p~tocol).

1791 1792 1793 1794 1795 17

1797 1798 1799

1800 1801

1$0~

1811

1$1~ 1813 1814

1815

le fi3. Bracketing design and?,hree-fourthsMatrix (over$l si=: o n e - ~ f of full testing protocoi)

1816

1817 1818

4. Data ~ v a l ~ t i o n

The stabilitydata obtained under a~ a ~ i protocol n g should besubje~tedto the sametype

1820 1821

1825 1826 1827 1828 1829 1830 1831 1832 1833 1835

p r o d ~ ~ sites o n and the lengthof s t a b ~ ldata i ~ on these~ ~ h efors , ANI", BLA or PLA applicatio~Applicants are advised to consult c h e review ~ team when queitions arise. 2. ~

gBLAs, or PLAs ~

In p ~ c i p l eprimary , stability batches shouldbe made at the 1839

1843 1844 1845

site is equivalent. If at thet h e of application data and6 months of accelerated data on three the intended commercialsite, a reduced numbe d ~ t i of o data ~ than the primary batches may be acceptable.In ~ ~ ~these o site-specific n y batches may be of pilot scale.

1846 1847 1848 1849 1850 1851 I852

A drug substance shouldbe adequately c h ~ a c t e r(i.e., i ~ results of chemical, physical, be~of andy when applicabley biological testing). Material produced at ~ ~sites should ~ c ~ m p ~ bqudity. le In general, three tosix months of s~bilitydata on one tothree sitespecific drug substance batches, depending on theav~labilityof s ~ c i e nprimary t stability data %om another site, should be provided at the time of ap~licationsubmission. Table 1I depicts the site-specific stability data recommended for thedrug substance in an originai applicatio~ 58

1853 1854 1855

Scenario"

1856 1857 1858

S u ~ ~ i eprimary nt stabiti~ data are available for the drug substance

1859

S ~ l c i e nprimary t stabiliQ data are not available for the drug sub^^

1862 1863 1864 1867 1868 1869

1870 1871 1874 1875 1876 1877

"%e phrase suflcient primary stability data means that,at the timeof submission, therearc 6 months of ac~~lmted data and at least 12 months of long-term data onthreeprimary stability batches made at di&rcnt a pilot or production site &om the intendedsite. Additio~~ long-term stability data an^ ifapplicable, accelerated data, should be submitted for review as soon 8s they become available prior to the approval. A ~ m m ishould ~ ~be provided t in the applicationto place the first three productionbatches at each site on long-term and acceleratedstability studies and annual batches thereafter on long-term studies using the approved protocol and to reportthe resulting datain annual reports.

The complexityof the drug product dosage formis a critical factorin determining the number of site-specific batches foran original application. The quality andlor stability o f a simple dosage formis less likely to vary due toa digerentm a n u ~ site ~ ~,~ g ofa complex dosage form. Three site-specific batches are needed fora complex dos ~ l profile for the product made at provide an independent and statistically r n ~ gstability that site. One site-specific batch maybe sficient to verify the stability profileof a simple ~ s stability data recommended for dosage form. Table 12, below, i I l ~the~ site-specific drug products in an original application:

1878 1879

1881

The phrase ~

data and at ~ u ~ i o n

1896 1897

~primury e nsrubilily t &tu means that, atthe time of ~ ths o f long-termdata on three p site.intended

~ stability m b~

ithere~are 6 imonths o o f~ a ~ l ~ t ~ d a made~at a d~ i ~ ~pilot n tor

Additional long-term stability data and, i f applicable, accelerated become av~lableprior to the approval.

mi@& for review as soon as they

A mmitment should beprovided in theapplication to placethebatches.at and a ~ l c ~ t stability ed studiesand annuai batches ~ e ~on ~ e r to report the resulting datain annual repom.

each site on long-tm using the approved protoMI and

1898 Other factors, suchas lack o f experience at the new site in a ~~c~~ dosage form, or 1839 * the e n v i ~ ~ e n tconditions a l between ally the 1 ~ 0 stability r product. drug o fTherefore, a o eatch 1901 s ~ x c iin~these t cases. More than one site-specific batchmay be needed for a drug X 902 s u ~ ~ ~ p r o dthat u cist intrinsically unstable. though one site-specific batch maybe sufficient undercertains i ~ o n sthe , data so

I903 19~4 1905 1906 1907 1908

particularly i f limited toaccel r the establishmentof a retest single site-specific batch mayonly serve ~ u b ~ c ~ p r o dthat u chas t been established based on p~~~ s ~ b i lbatches i~ at a pilot

1909 1910 1911

In general, site-specific drug product batches shouldbe made with identifiable site-s~cific &g substance batches both for original applications, wherever~ s s i b l eand , for postapproval stability commitment.

1912

Although pilot and commercial facilities may or may not be located on the same campus or

1913 1914 1915

1920 1922

ific Data ~ a c ~ g e

~ ~ for ~ ANLh4.s e n & t i o ~

1925 1926 1927 9 ~ $ 1929

For ANDAS, the primary batch(& to support the application are usually in the p~oduc~on facility. If thestabilitybatch(es) mnotmade at the intended c o ~ e r csite, i ~ stability data shouldbe generaw as 0~~~ & Table 13,on thedrug product ~ u f a c at~that d site,i.e. site-specific batches, and the data should be i n c l u d ~ e the product made at each site is equivalent. in the original sub~ssionto d e m o ~ t that

1930 1931 1932 1933 1934

If the pilot plant where the primmy stability batches are made is located at thei n ~ ~ d e d commercial site (i.e., on the same campus as the intended commercialfacilityj, the site-specific stability~ c o ~ e n & t are i omet ~ and no additional data willbe needed. A comxni~entshould be made to place thefirst three production batches andannual batches thereafier on long-term stability studies.

1938 1939

For complex dosage f m s as described in the previous section, a reduced number of site-specific batches may be justifiedif accelerated and long-term data are a ~ ~ l a batl ethe time of appli~tion ~bmission on batches made at diserent a pilot or comxnexiaf siteh m the intended commercial facility.

Tabie 13: Site-Specific S t a b i i i ~ Data for a Drug ~ r o ~ u in c tan ~ r i g i ~Aa i ~

1940

1941 First 3 p r ~ u c t i o n ~ t con hes long-term ~ b i lstudies, i ~

1942 1943

First 3 pr~uctionbatches on a

1951 1952

Drug substancebatches used to produce site-specificdmg product batches should be ~ 1 w l identified y Additid

should be submitted for review as soon ac they become availabia prior to approvat. A commitment should be provided in the applicationto place thefm three pduction batches at each site on longterm stability studies and annual batches thereafter on long-term studies usingthe approved pmtocol andto report the resulting data in annual reports.

J. ~ h o t o s t a b i i i ~ 1. General

1953 1954 1955

The lCH H a r ~ o n ~ e d T r Guideline ~ a r ~ t e on Stabili~Testing of New Drug Substances and Products (hereafter referred toas the parent ~ i d a n c enotes ) that light testing should be an integral parto f stress testing.

1956 1957 1958 1959 1960 1961 1962 1 1 1965 1966 1967 1968 1969

The ICH QlB guidance Photostabili~Testing of New Drug Substances and Products primarily addresses the generationo f photos~bility~ o ~ t i for o n new ~ o l ~entities u l and ~ associated drug products and the useof the data in determining whether p r ~ u t i mo w ~ e s in ~ u f a ~ nlabeliig, g , or packaging are needed to tig gate exposure to light. QIB does not specficaliy address other p h o t o s ~studies ~~i~ that may needed to support, for e m p i e , the photostabiIity o f a product under in-use conditionsor thep h o ~ s ~ b i l iof t y analytical samples. Btzeause data are generated on a directly exposed drug substance alone a d o r in simple solutions and drug products when studies are conducted as described in the Q 1B guidance, ~ o ~ i e d of ge p ~ ~ ~ characteristics b i l i ~ maybeusefkl in d e ~ d e~n additional ~ g studies may be needed or in providing justification for not performinga d d i t i o ~studies. For example, i f a product has been determined tophotodeg~deupon direct exposure butis adequately protected by packaging, an in-use study may be needed to support the use of the product (e.g., aparente~ldrug that is i d b e d over a periodo f time). The test conditions for in-use studies will vary depend~gon the product and use but should depend on and relate to the drections for useof the particular product.

1970 ~9?1 1972

~hotostabilitystudies are usually conducted only in conjunction with the first approvalof a new molecular entity. Under some circumstances, p h o t o s ~ b i lstudies i~ should be repeated if certain postapproval orsupple men^^ changes, such as changes in f o ~ u I a t i or o ~packagi~g,are made to 62

1973 1974 1975 I976 1977 1978

the product, or i f a new dosage form is proposed. Whether these studies shouldbe repeated depends on thephotostabili~characteristics determined at the h e of initial filing and thetype o f changes made. For example,if initial studies demonstrate thatan active moietyin a simple solution degrades upon exposureto light and the tablet drug productis stable, as ~ s ~f i l~i i n t the ~ u e s t i n gagproval of aliquiddosageformmaywarrant ~ ~ t istudies o ~tol stics o f the new dosage form.

1979 1980 1981 1982 1983 1984

P h o t o ~ b i lstudies i~ need notbe conducted for products that duplicate a commercially available listed drug product provided that the packaging ( i i e d i a t e containwklosure and market pack) and iabeIing storage statements regardmg light duplicate those o f the referencelisted drug. If deviations in packaging or labeling statementsare d e , additional studies maybe ~ ~ e n d e d . The d ~ i s i o nas to ~ e t additional h ~ studies shouldbe conducted willbe made on a case-by-case basis by the chemistry review team.

1985 1986 1987 1988 1989 1990 1991 1992

‘he intrinsic p h o t o ~ bcharacteristics ~i~ o f new drug substances and products should be evaluated to d e m o ~ t tehat,as appropriate, light exposure doesbnot result in ~ c c e p t a b l e change. Normally, p h o ~ ~ b i l itesting t y is camied out on asingle batcho f material selectedas described in the section Selectionof Batches, in the parent guidance. Under somec ~ c ~ ~ c e s , these studies should be repeated i f certain variations and changesare made to the product (e.g., formulation, p a c ~ g ~ g )Whether . these studies should berepeated depends on the photostability characteristics d e t e ~ e at d the timeof initial filiig and the typeo f variation (ind/or change made. CiCH QlW

1993 1994

A systematic approach to photostabili~testing is recommended covering, as appropriate, studies such &S:

1995 1996 1997 1998

Tests on the drug substance; Tests on the exposed drug product outside of the immediate pack and if necessary, Tests on the drug product in the immediate pack;and i f necessary, Tests on the drug product in the marketing pack.pCH QlB]

1999 2000 2001 2002

The extent o f drug product testing shouldbe established by assessing whether or not acceptable change has occurred at the end o f the light exposure testingas described in Figure 2, the Decision Flow Chart for Photostabili~Testing o f Drug Products. Acceptable change is change within l i i t s justified by the applicant. DCH QlB]

2003 2004

The formal labelimg requirements for photolabile drug substances and drug products are established by nafional/regional requirements. [rCH QlB]

2005 2006 2007 2008 2009 2010 201 1

2. Light Sources The light sources described below may be used for photostabili~testing. The applicant should either maintainan appropriate controlo f temperature to minimize theeffect o f localized temperature changes or include adark control in the same environment unless otherwise justified. can rely on the spectral For both options 1 and 2, a pharmaceutical m~ufac~er/applicant distribution specification of the light source manufacturer. PCH QlB] 63

2012

201 3 20 14 201 5 2016

Any light source thatis designed to produce an outputsimilar to the D65AD65 emission standard such as an ~ f i c i a l ~ y l ifluorescent ght lamp ~ m ~ visible ~ nandg violet 0outpttts, xenon, or metal halide lamp. D65 is ly recognized standard for o u ~ ~ r ~ as y l i ~ h t defhed in IS0 10977 (1 993). ID65 indoor indirect ~ y l i g hs~~ t For a source emitting significant radiation below320 ~ o m e t ( em ~) an ,a p p ~ p filter@) ~ a ~my fitted to eliminate such radiation. ~ ~ H Q l ~ ]

201 2021 022 2023 2024 2025

For option 2 the same sample~ ultraviolet lamp. 0

o be exposed ~ d to both thecool white fluorescentand

A cool white ~ u o ~lamp ~ designed n t toprodue an output similar to that 10977 (1993); and A near W fluorescent lamp having aspectral distribution from 320 m to 400 m witb a maximum energy emission between350 nm and 370 nm; a s i ~p r ~o ~ ~ ~ oof W nt should be in both bands of 320 to 360 m and 360 to 400 nm. UGH QIB] 3. Procedure

FCH QIB]

~029 2~30 203 1 2032

For c o ~ a t studies, o ~ samples shouldbe exposed to light p r o ~ an ~ overall g i ~ ~ t i ofo n not less than l .2 million lux hours and an integrated near u l ~ v i o l eenergy t of not less thax~200 watt hodsquare meter to allow direct comparisons to be made between the drug substance and drug product.

2033 2034 2~35 ~036

Samples may be exposed side-by-side with a validated chemical actinometric sydem to ensure the specified light exposure is obtained, or for the appropriate durationof time when conditions been monitored using calibrated radiometers/lux meters.An e m p l e of an ~ c t i n o m e procedure ~c is provided in the Anne%

~ 0 3 I~f protected samples(e.g., wrapped in a l ~ foil) n are~usedas dark controls to evaluate the ~ 0 3 8 ~ n ~ b o f~thermally o n induced change to the totaI observed change, these shouldbe placed 9 alongside theau~enticsample. (JCH QlB]

4. Drug Substance [ICH QlB]

2040

2041 2042

For drug substances,photostabili~testing should consisto f two parts: Forced degradation testing and codkmatory testing.

2050 205 l 2052

The purposeof forced degradatio~testing studies is to evaluate the overallphotosensitiv~~ o f the for method development purposes andlor pathway elucidation. This testing may involve the drug substance alone andlorin simple s o l u t i o ~ s u s ~to~ validate i o ~ the be in chemically inert and ~ ~ n tical procedtires. In these studies, the samples should containers. In these f o d d e ~ ~ studies, t i o na variety of exposure conditions maybe used, of light depending on thep h o t o ~ ~of~thet drug i ~ ~sub^^ involved and theintensi~ the sources used. For developmentand v ~ i ~ t i purposes, on it is a p p ~ p ~ atot el i i t exposure and end the studies i f extensive d~mpositionoccurs. For photostable materials, studies may be terminated &er an appropriate exposure level has beenused. The design of these experiments is left to the~ p l i ~ t discretion 's thou^ the exposure levels w d should bej ~ ~ ~

2053 2054 2055 056

Under forcing conditions,decom~sitionproducts may be observed that are d i e l y to be formed This ~ o ~ t i may o nbe useMin developing under the conditions used for c o ~ ~ a t ostudies. ry and validating suitable analytical methods. If in practice it has been demo^^ they are not f o ~ e in d the c o ~ ~ astudies? t o ~ thesed e ~ ~ t i products on need notbe examined further.

2057 2058 2059

~ n f studies ~ should a then ~ be u~n d e ~ e to n provide the ~ o ~ a t i necessary on for handling, packaging, and labeling (see Section VIII.J.3., Procedure, and 4.a., Presentation of Samples, for ix&omtion on the designof these studies).

2060 206l 2062 2063' 2064

Normally, only one batchof drug substanceis tested during the development phase, and then the as d,escriW in the p ~ o t o ~ b characteristics ili~ shouldbe confirm& on a single batch selected parent guidanceif the drug is clearly photostable or photolabile. I f the results ofthe confirmatory study are equivocal, testing of up to two additional batches shouldbe conducted. Samples should be selected as described in the parent guidanw.

~~~

2065

a. Presentation of Samples UCH QlB]

2067 2068 2069 2070 207 1 2072

Care should be takento ensure that the physical characteristicsof the samples under test are taken into accoung and efforts shouldbe made, such exs cooling and/or placing the samples in sealed containers, to ensure that the effects o f the changes in physical states such as sublimation, evaporation, or melting aremini-. All such precautions should be chosento provide minimal intederence with the exposureof samples under test. Possible interactions between the samples and any material usedfor containers or for general protection o f the sample shouldalso considered andelidnated wherever not relevant to the test being canied ouk

2073

A s a directchal~en~e for samples of solid drugsubstances, an appropriatemount oE sample should

2074 2075

be taken and placed in a suitableglass or plasticdish and protected with a suitable transparent cover if considered necessary. Solid drug substances should be spread across the containerto give 66

2076 2077 2078 2079 2080 208 1

a thicknesso f typically not more than 3 millimeters. Drug substances that are tiquids should be exposed in chemi~llyinert and~ p ~containers. n t b. Analysis o f Samples

be examined for any changesin physical At the endo f the exposure period, the samples should e, clarity or color o f soiution) and for assay and degradants by a method likely to arise&om p h o t o c h e ~ ~ ~processes.' g~tion

2082 2083 2084 2085 2086

are involved, ~ p ~ve should i g .substance samples i n d i v i d ~tests. Similar sampliig ~ ~ i d e ~ t i o ~ , of the e, apply to otherm a t e r i ~that may notbe h o m o ~ e after n ~ ~exposum. The analysis sed sample shouldbe performedc o n c o ~ with ~ ~thyat of any protected samplesused these are used in the test.

2087

c. Judgment o f Results

2088 2089 2090 209 1 2092

The forcedd e ~ ~ t i studies o n shouldbe designed to provide suitableor mat ion to develop and validate test methods for the ~ ~ astiddies. t These o test ~ methods should be capableof resolving and d e t ~ n photolytic g degradants that appear during the confirmatory studies. Men e v a l ~ the ~ gresults o f these studies, it is important to recog& that they form part of the stress not therefore designed to establiih q ~ i ~ t i or v qe ~ t i t a t i v limits e for change.

2093 2094 2095 2096 2097

The ~ ~ ~studies a should t identifv o p~ ~ u t i measures o ~ needed in m a n ~ a orcin~ f o ~ u l a t i o ~the odrug f product andi f light ,resistantp a c ~ g i n gis needed. When evaluating the t o ~to determine whether change due to exposure to light is acceptable, results of c o ~ ~ astudies it is important to consider the resultsfrom other formal stability studies to ensure that the drug will be within jusW1ed liits at time of use (see the relevantICH stability and impurity guidance).

2098 2099 2100 2101

5. Drug Product @CH QlB]

2102

2103 2104

~ o ~ itheystudies , on drug products should be carried out in a sequential manners ~ with~ testing thefilly exposed product then progressing as necessary to the product in the immediate pack and thenin the marketing pack. Testing should progress until the results demonstrate that the drug product is adequately protected &om exposure to light. The drug product should be exposed to the light conditions described under the procedure in Section VIIJ.3,

2105 2106 2107 2108

Normally, only one batcho f drug product is tested during the development phase, and then the be c o ~ ~ one adsingle batch selectedas described in the p h o t o ~ b i lci ~h ~ c t eshould ~ i ~ parent g u i d ~ c e the i f product is clearly photostable or photolabile. Ifthe results ofthe ~ ~ ~studyaare equivocal, t o testing ~ of up to two additional batches shouldbe conducted.

2109 21 10 2111

For some products whereit has been demons~tedthat the immediate packis completely impenetrable to light, suchas aluminum tubes or cans, testing should normally only be conducted on directly exposed drug product.

2112 21 X3 21 14

It may be approp~ateto test certain products, such as idision liquids or dermal creams, to support their p h o t o s ~ b i lin-use. i~ The extento f this Wing should depend on and relateto the ~ r e c t i for o ~use,and is left to the applicant's discretion.

21 15 21 16

21 17 21 l8 21 19 3120

2124 2125 126 127 212 2129

313~ 2131 2132 2133

2134 2135 2136

2141 2142

2 145

2146

"%em practicable when testing samples be presented in a way similar to the sho~d

o f the primary pack$ these

ii for the drug substance. The to the light source. For

is not practical (e.g., dueto oxidation o f a product), the sample should be placed tective inert~ p ~conthiner n t (e.g., quartz). If testing of the drug productintheimmediatecontainer or as is needed,thesamples l yrespect to the lightsource, whichever provides should be placed h o ~ n ~orl ~y v e ~ ewith for the m uniform exposureof the samples. Some adjustment o f testing conditions may have to testing large volume containers (e.g., d i s ~ packs). ~ ~ g

b. Analysis o f Samples At the end of the exposure period, the samples should be examined forany changes in physic^

properties (e.g., appearance, clarity, or coloro f solution, d ~ s o l u t i o ~ d i s i n t e ~ tfor i o ndosage such as capsules) asuifor assay andd e g ~ ~byt asmethod suitably validated forp ~ d u c ~ to & s e~ ~ ~ ~ h o t o c h e ~ c a l processes. de~~tion When powder samples are involved, sampling should ensure that arep~sentativeportion is used in ~ d i tests. ~ For d solid ~ oral dosage form products, testing shouldbe conducted on an

appropriately siEd composite of, for example, 20 tablets or capsules. Similar sampiiig considerations, suchas h o m o g e ~ ~ t i oornsolubil~tionof the entire sample, apply to other ~ t e that~ may a not~be homogeneo~af'ter exposure (e.g., creams, o i n ~ e ns~pensions). ~, The c o n ~ ~ ~with t lthat y of any p ~ t ~ t wlysis of the exposed sample should be pe&ormed samples used as dark controls if these areused in the test.

c. Judgment o f Results 68

2147 2148 2149 2150

2151

2152

Depending on the extentof change, special labelii or packaging may be-neededto mitigate e exposure to light. M e n evaluating the resultsof photostability studies to d e t e ~ whether change due to exposureto light is acceptable, itis important to consider theresults obtdned from other f o d stability studies to ensure that the product will be within proposed s ~ i ~ ~ t i o ~ the shelf life (see the ~ l e vICH ~ t ~ b i i and i ~ i m p ~ t yg ~ ~ ~ ) . 6. ~~e ~e~~

153 2154 155 156

2157 2158 2159 2160 2161 2162 2163 2164 2165 166 2167 2168 2169

Technolo system, the same approachmay be used,but a the lightsource used. Prepare a sufficient quantity of a 2 percent weigh~volumeaqueous solutionof q ~ l o ~ ~ h y ~dhydrate c ~ o (iinwessaxy, r i ~ dmlve by heating).

Put X0 ~ 1 l i l (id~ ., of )the solutioninto a 20 mL colorless ampoule (see drawn i g,below), h e ~ e t i ~and y , use this as the sample. Separately, put 10 mL of the solution into a 20 xnL colorless ampoule(seenote X) hermetically,wrap in a l foil to protect ~ cornple ~ from light, and use this as the ,Exposethesampleandcontrol to thelight source fot a p ~ o p ~number te of hours. M e r exposure, determine the absorbances e r path length. Calcd the control (AO)at 400 nm using a 1~ n ~ e t(cm) absorb~ceunits (AU): A = AT AO. The length of exposure should be s ~ c i e ntot emu change in absorbance of at least 0.9 AU. Note: Shape and ~ ~ ~(SeeeJapanese ~ Ii n d ~o ~ ~ a (JIg n R3512 ~ d (1974)for ampoule s p e c ~ c a t i o ~ ) . "

Yoshioka, S., "Quinine Actinome~as a Method for Calibrating UltravioletRadiation Intensityin Lightstability Testing of Pharmaceuticals,"Drug Developmentand Industrial Pharmacy,20( 13):2049-2062, I994 69

ption 2

2170 2171 2172 21173 211~4 2175 2176

FiS a 1 c m quartz cell and use this m the sample. ~ e p ~ t efill l ya l cm quartz cell, w p in foil to protect ~mpletely this as the control. Expose the sample and control to the iight source for an approp~ate f hours. Afterexposure,

2177 2178

Alternative ~ c k a g i n g c o ~ g ~ may t i o nbes used if appropriately val chemical ~ ~ o m e t e be ~ used.m y

~~~

~ ~ r b o fathensample ~ (AT) and the absorbance, )A = AT AO. The lengthof exposure abso~anceof at least 0.5.

2179

2380

2181

2183

2184 2185 21 86 2187 2188 2189 190 191 2192 2193 2194 2195 2196 197 198 2199

The extento f the drug productphotostabili~testing dependsora the change that has occurred at 2, above, theDecisih Flow Chart for ~ h o ~ ~ b the egdo f each test tier described in Figure a ~ p ~criteria c eforthe Testing of DrugProducts. Test results that are outsidethe product would notbe considered acceptable change. This intrinsic p h o t o ~ b i icharacteristics i~ ofnew drug substances and products,+d no correlation has n developed to equate within a s p ~ i f i ~ t i oresult n to an expirationd a ~ period. g The a ~ p t a ~oflany i ~observed changes shouldbe justified in theappli~tion. It m y be ~ p oto ~ t consider other degradative processes (e.g., thermal) when j ~ ~ i a nphotos~bility g change as acceptable because the processesm y be dependent and ~ ~ t i v For e . example, a5 percent loss in potency dueto p ~ o t o d ~ g ~ ~may t i obe n considered acceptable if that is the only type of degradation observed. overrmnt I f the the she~f-~fe due ptodegradation, thermal the ~ a the ~ e p ton a bbased le potential additive effectof the changes. In this case, precautions shouldbe taked to~itigatethe product's exposure to light.

Under the intense light exposure conditions included in the Q 1B guidance, certain colorsin solid dosage € o m may fade. Q ~ t i ~ analysis v e o f the color changeis not r e c o ~ e n as d~ these changes m not lihly tooccur under actual storage conditions.In the absence of change in other may be a ~ p ~ b i e . parameters suchas assay, these color changes 8. Photostability Labeliig Considerations

The data generated using the procedure described in the ICH Q1A guidance is useh1 in determining when special ~ n d l i n gor storage statements regarding exposure to light should be included in the product labeling(21 CFR 201.5~~)(4)).The labeliig guidance provided below ~~3 o ~s ~ s b i l i ~ that ~ t e ~ e n ~ pertains only to productsas packaged for d i ~ b u t i o nI, ~ ~ c t iand 2204 to 21 CFR 201.57(j) are not covered. pursuant ~ ~ 0 5 may be needed to address in-usec o n d ~ t i o

22~2

2206

Change after direct exposure: If changes that are observed when the productis directly exposed 70

2207 2208

under the light conditions described in the Q 1B guidance are acceptable, no labeling storage statement ~ g ~ d i light n g is needed.

2209 2210 2 2 11 12 2213

~ ~ a after n ~~ xe ~ Q §inut r ~~immediate e container/clQ§ur~:I f changes observed when the

2214 2215 2216 2217 2218

2219

2220

2223 2224 2225 2226 2228

m 9 2230 2231 2232 2233 2234 5 2236 2237 2238

2239 2240 2241

2242 2243

For those products that are d i l y

ved h m the ate container, such as to the patient, ando p h ~p ~ c u c t sthe ,

For products that maybe removed f b m the immediate pack, suchas pharmacy bulk packs, a ~ t e m eshould ~ t be included suchas “PROTECT FROM LIGHT. Di l i ~ t - ~ ~ t ~ n ~ e r . ~ ~ ~ a after u ~e x e~ Q § uin.r ~ the m ~ r ~pack e t If changes that are observed are acceptable only when the product inthe marlcet pack is exposed under the conditions described in the QlS guidanw, l a ~ l i n gstorage ~ t e m ~regarding t s light shouldbe included. Examples of typical storage statements are, for single-dose and multip1e~ose.products r e ~ c t i v e l y“PROTECT , FROM LIGHT. Retain in carton until time o f use.” and ‘*PROECT FROM LIGHT. Retain in carton until contentsare used.”

When d e g ~ ~ t i products on aredetected upon storage, the followingi ~ o ~ t i about o n them should be submittd Procedure for isolation and purification Identity and chemical &-uctures Degradation p a t h ~ y s Physical and chemid properties Detection and q ~ t i ~ t i levels on Acceptance Criteria~ ~ d iand ~ total) d ~ l Test methods V a l i ~ t i o ndata Biological eEect and p h ~ a ~ l o g i cactions, al including toxicity studies, at the con~n~tio~ likely to be e n ~ ~ t e r (cross-reference ed toany available ~ o ~ t i iso~nc e p ~ b l e )

If~ e r n i ~ o~f the o ndrug substance in the dosage form is possible, the~ o ~ a t i descri on above also should be provided.

L Thermal ~ y c l i n ~ A study of the effectso f temperature variation, particularlyi f appropriate for theshipping and 71

storage conditionso f certain drug products, should be considered. Drug products sus~ptibieto phase s e p ~ t i loss o ~ o f viscosity, p ~ c i p i ~ t i oand n , agg~gationshould be evaluated under such be cycled thermal conditions. A s part o f the stress testing, the packaged drug product should through ~ m p conditions e ~ that ~ simulate the changes likely be to encounte product is in ~ s ~ b u t i o n .

A ~ r n p cycling e ~ study ~ for drug products that may be above freezing m y consist of three cycles of two days at re followed bytwo days under ~~e~~ storage con~tions(40"~).

.Atern^^^ t e m p e cycliig ~ ~ studyfordru may consist o f

mture (-10" to "20°C)

followed by two days under

5 6 7 2258 2259 2260

2261 62

For

on aerosols, the~~~d~

etween ~ u b ~ ~ g

cycle study consists of or €our six-hourcycles t e (75-85 r n percent ~ m RH ~)for a period o f up to and 40°C

.For h = n drug products, the ended cycle study should include e v ~ ~ ofo eE&s n due to accelerated thawingin a microwave or a hot water bath c o n ~ din ithe~ ~ unless

hbelii.

Alternatives tothese c o n ~ t i may o ~ be acceptable with a p p ~ p r i a ~

j~tifi~tion~

M.~ ~ b i l ~ in~ Forei T ~ t i n ~

2263 2264 2265 66

S ~ b i ltesting i ~ (aswell 8s fnished product release testing) pe~ormedin any foreign or domestic facility may be used as the basis for approvalo f an appli~tion.Tiis ~ ~ l udI~"eS, s ANDAS, and related Ch4C supplements. A satisfactory inspectiono f the l a ~ r a ~ r ythat ~ ewill s ~perform the testingwill be neCessmy."

2267 2268 2269

Applicants should consider the effects o f bulk p a c ~ gshippingy , and holdingo f dosage forms and subsequent market packaging, and d i ~ b u t i o n the o f finished drug product, andbe aware of

e dthe a operation shouldbe es~blishedand s u b s ~ ~ a t by

"This statement replaces a previous position, established via a CDER Ofice o f Generic h g s guidance, which r ~ m m e n that d ~finished product andstability testing be conducted at a United States i a ~ forodnrg~ productsr n ~ u f inaforeign ~ ~ facilities andshipped in bulk ~ t a i Mnthe~ United States for ~ ~ g i into n g immediate containers for marketing.

esting of ~ i o t e c h n o l o ~

2274 2275 6 2277 227$

~~$

2289 2290 2291 2292 2293 2294 95 96 2297 2298

2300 2301 2302 2303 2304 2305

231 1 2312

1. General @CH QSC]

The ICH harmonized tripartite guidance entitled Q1A S t ~ b i l Testing i~ of New Drug Substances and~ ~ o d u cissued ts by ICH on October 27,1993, applies in general tob i o t ~ ~ o l o g i ~ i o l o g i ~ l p r ~ u c t s . H o w ~ e r , b i o t e c ~ o l o g i ~ products i o i o g i ~have l distinguishing c ~ c t e ~to~ c s

The ~ b i l i ~ nm y ~~ m p l eianalytical x 2ethodologies.. ~ Assays for biol ,whereapplicable,should be part o f thepivotit1stability studies. Approp~ate p h y ~ i ~biochemical, h e ~ ~ and ~ ~ o methods c for the h analysis ~ of the~mol^^^ entity and theq~titative de~ctiondegradation of products shouldalso be part of the stability o f product permit useof these program whenever purityand molecular c ~ c t e r i s t i c s the ~ethodoiogies. With these ~ n c inemind, ~ theapplicantshoulddevelopthepropersupporting S duct and consider mmy external~ n ~ tthat i can o ~

for a

guidance toapplican~r e g ~ n the g typeo f ~ b i i istudies ty that shouldbe provi&d in sup^^ of m ~ appli~tions, ~ eIt is ~derstood ~ that ~ during the review and evaluation process,con^^^ updates o f initial stability data mayowur. 2.

cope @CH QSC]

The g~~~ in this section applies tow e l l ~ ~ cproteins t e ~and ~ which are d ~ ~ tandi products v ~ o f which they are~ m p o n e nand using r e ~ m b ~ t d ~ ~ ~acid ~ (~-DNA) n u c l t e~i ~c o l o Thus, the section covers thegene~tionand submission o f stability datafor products suchas c ~ o ~ n e s ( i n t e ~ e r o ~ , colony-stimulating interle~, factors, tumor necrosis factors), e ~ o ~ ip l ~e i n~o g ,e nactivators, blood plasma factors, growthh o ~ o n e and s growth o ~ and vaccines consistingof w e l l ~ ~ cp rt oet e~~ ~ factors, i ~ u l i n sm , o n ~ l antibodies, polypeptides. In addition, the guidance outlinedin the following sections may apply to other ty of products, such as conventio~lvaccines, afterc o ~ u l ~ t i owith n the product reviewofice. The blood, or cellular section does not cover~tibiotics,allergenic extracts, heparins, vitamins, whole blood components. 73

2313 2314 23 15 23 16 317 231

3. Terminology [ICH QSC]

For the basic terms used in this section, the readeris referred to theGlossary, However, because m ~ u f a c ~orf s b i o t e c ~ o l o g i ~ ~ i o products l o g i ~ l sometimes use~ d i t ~ o n~a l n o ~ o ~ , traditional t e r n are specified in parentheses toassist the reader. 4. Selection o f Batches [EHQSC] a Drug Substance @ u l k ater rial)

2328

W e r e bulk materialis to be stored after ~ u f i a cbut ~ before , fo~ulationand final m ~ u f a c stabiiity ~ n ~ data should be provided on least at threebatches for which and storageare r e p ~ e n ~ovf the e ~ u f i scale ~ o fgproduction. A o f six on^ ~ ~data at~the tirne t of ~ y b ~ ~should i o ben~b~~ in cases where storage periods greater than six months are requested. For drug substanceswith stomge periods of less than six months, the ~m~ mount o f stability data in the initial submission shouldbe d e t e on~ a ~ dr d dd e of case-by-case basis. Datafbm pilot-scale batchesof drug subs^^ p ~ d u at~ ae f e ~ e n ~ t i and o n purification maybe provided at the time the ~ ~ l i ~ tisi submitted on to the Agency with a wmmitment to place the first three~ ~ a scale cbatches ~ intogthe long-t~ stabirity program after approval.

2329 2330 233l 2332 2333 2334 2335 336 2337

The quality o f the batcheso f drug substance placedinto the stabilityp r o ~ , s h o u l be d represen~tiveof the quality of the material used in precliical and cliical studies andof the quality o f the material tobe made atm ~ u f a c scale. ~ g In addition, the drug substance (bulk ma~rid) made at pilot-scale should be produced by a process and stored under conditions~ p ~ ~ ~of t i v e ~ scale. The~drug substance ~ intothestability g programshould that used for the m be storedin containers that properly represent the actud holdmg c o n ~ e rused s during m ~ u f i a cContainers ~. of reduced size may be acceptable for drug substance stability testing provided that they are constructedo f the same material anduse the same type of+ontainer/closure system that is intended tobe used during manufacture.

~~~~

2323 2324 2325

2335 2339

2344 2345 2346 2347 2348 2349

b.Intermediates

During m ~ u f a ocf ~ ~ i o t e c ~ o l o g ~ ~ i products, o i o g i c the ~ quality andcontrol of certain i n t e ~ e ~may a ~be s critical to the productionofthe final product. In general, the ~ u ~ ~ s generate in-house dataand p m s s liits that ensure their should id en ti^ i n t e ~ e d i a and stability ~~n the bounds of the developed process. AJthough the use of pilot-scale data is ~ ~ s ~ i bthel manufiiturer e , should establish the ~ t a ~o f such l i data ~ using them ~ ~ a c ~ g scale process. c. Drug Product (Final Container Product)

Stability ~ o ~ t i should o n be provided on atleast three batchesof final con~inerproduct representative of that which wilt be used at m ~ u f a c ~ scale. n g Mere possible, batchesof fmal from different batchesof bulk container product included in stability testing should be derived material. A minimum of six ~onths'data at the time of submission should be submitted in cases 74

r a ~Notfor I ~ i ~ ~ n t a t i o n 23 50 51 2352 2353 354 2355 2356

where storage periods greater than six months are requested.For drug products with storage periods o f less than six months, them ~ amount m o f stability ~ data in the initial submission should be determined on a case-by-case basis. Product expiration dating should be based upon the actual data submitted in support of the,appli~tion.Because dating is based upon the ~ - ~ ~ x a l - ~ m pdata e rsubmitted a ~ e for reviews continuing updates o f initial stability data should occur during the review and evaluation process.The quality ofthe final container product pl& on stability studies should bepresentative ofthe qualityo f the material 4in the ’

2364

2365

o f material usedin pmlidcaI and clideal studies, the the dating periodor is not r e p ~ n t a t i v e the a p p l i ~should t notify the appropriateFDA reviewing office to determine a suitable courseof action. d.Sample Selection

2366 2367 2368 2369

%%ex one productis distributed in batches differingin fdl volume (e.g., 1milliliter (d) 2, mL,or 10 mL),unitage (e.g., 10 units, 20 units, or 50 units), or mass (e.g., l milligram (mg), 2 mg, or S me), samples tobe entered into the stability program may be selected on the basis of matrix a system and/or by bracketing.

2370 237 l 2372 2373 2374 2375 2376 2377

hctions of samples are Matrixing the statistical designof a stability study in which different doc~entationis tested at different sampling points should only be applied when appropriate provided that confims that the stabilityo f the samples tested representsthe stability of a l l samples. drug product should be identifieda s ,for example, The differences in the samples for the same covering dflerent batches, different strengths,Werent sizes o f the same closure, and, possibly,in some cases, different c o n ~ e r / c l o ssystems. ~e ~ a should~notbe applied% g samples with diffexnces that may s i t stability, such as different strengths and different containersklosures, where it cannot beconfiied that the products respond similarly under storage conditions,

2378 2379 23 80 2381 2382 2383

W e r e the same strength andexact con~er/closuresystem is used for three or more fill contents, the ~ u f ~may~elect e torplace only the s d l e s t and largest containersize into the program (i.e., bracke~g).The designof a protocol that incorporates bracketing assumes that the stability of the intermediate condition samplesarerepresented by those at the extremes. In certain cases, data may be needed to demonstrate thatall samples are properly represented by data collected for the extremes.

23~5 2386 2387 23 88

5. S t a b i l i t y - I n ~ ~Profile t ~ g UCH Q5C]

On the whole, thereis no single stability-indi~tingassay or parameter that profiles the stability i~1 Consequently, the manufacturer should characteristics of a b i o t e c ~ o l o g i c a ~ i o l o gproduct. propose a s~bility-indicatingprofile that provides assurance that changes in the identity, purity, and potency of the product will be detected.

23 ~ 319

At the timeof s u b ~ s s i o ~ , a p p shouid l i ~ ~have v ~ i d ~ t the e d methods that comprise the

s ~ b ~ l i ~ - i nprofile, d i ~ and t~~ thedata should be av~lablefor review, "ie ~ e t e ~ i ~ of a which ~on tests s h o ~ be. d ~cludedwill be p ~ u c t - s ~ i ~ c * to XIS intended am not be ~ 1 - i n c ls~h iov~ed, be ~ c ~ to dee ~ onn s ~~produc ~te a. P r o t o ~ l

2~94 239~ 2396

2400

b. Potency

2402 2403 ~4~ 5 6 2407 2408

When the use of a productis linked testing for potency shouldbe part of the stab ~ o d u ~ in this ~ e~ ~ r ipotency ~ d~ achieve its ~ ~ effect. n It is based d on~ the

2410

2412 2413

and the results should be reported in

~~~

~ reference o n material ~ should be ~ c l u ind the ~ assay.

2415 2416 241 7 2418 2419 2420 1

2423 ~ 4 ~ 4

2425

Purity and ~ o l e c u ~ ~

~h~cte~~tion

2426 2427 28 2429 2430 243 l

For the purposeof stability testing of the products described in this guidance? purityis a reIative t i other o n , h e ~ ~ g e n e i t i ethe s , absolute tern. Because of the effect of g l y ~ s y l a t i o ~ d ~ i d aor o g i ~is extremely difficultto determine. Thus,the purity purity of a b i o t ~ ~ o l o g i ~ i o lproduct

The degreeof p u r i ~as , well as the i n d i v i d ~and totalamounts of d e ~ p ~~d uof ~c the ~ o into thes ~ l lstu i d~ ies, should f acceptable d e g ~ ~ should o n be deri s u b h c e md drug productused in the

2437 243 8 2439

studies.

245 X 2452 2453

For substancesthat cannot be properlyc h ~ c t e rori ~products for whichan exact analysisof the purity cannot bed e t e ~ n e d through routine analytical methods, the applicant should propose and justify a l t e ~ t i t~ e e pro ~ ~ ~ d. Other ~roduct Ch~cteristics

54 2455 2456 245~

The

find ~ 0

2460 246 l 2462

follow^ product ch~acte~stics, though not specifically relating to

i o l o g i ~products l should be monitored and reportedfor the drug ~ ~ :

Visual a p o f the product ~ (color ~ and opacity ~ for solutio~suspensions;color, t e ~ e , and ~ssoIutiontime for powders), visible particdates in solutions orafter the r e ~ n s t i ~ t i oofn powders or lyophilized cakes, pH, and moisture level o f powders and lyophilized products. Sterility testing or ~ l t e ~ a t i v(e.g., e s ~ontainer/cios~e ~tegrity testing) should be p e r f o ~ e dat a initially and at the end of the proposed shelf life. ~~~

ppendix

2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 ~476 2477 2478 2479 80 2481 2482 2483 2484 2485 2486 2487 2488 2489 90 249I 2492 3 2494 2495 96 2497

Additives (e.g., stabilizers, preservatives) or excipients may degrade during the dating period of the drug product. I f there is any indication during prelim^ stability studies thatFction or degradationo f such materials adversely &ect the qualityof the drug product, these items may need to be monitored during the stability program. The ~ n ~ n e r / c lhas o sthe ~potentialtotheproductadversely evaluated (see below).

and should

6. Storage Conditions PCH QSC] a. Temperature

Because most finished b i o t e c ~ o l o ~ ~ i o l oproducts g i c ~ need precisely defined storage

tern^^^ the storage conditions for the r ~ - ~ ~ ~ - tstabdity e ~ studies ~ r may a be ~ e confined to the proposed storage temperature. b.Humidity B i o t ~ ~ o l o g i ~ i o lproducts o g i ~ are generally disbibuted in containem protecting them against humidity. Therefore, whereit can be demonstrated that the proposed containers (and conditions of storage) Sard sufficient protection against high andlow humidity7 stabilityfests at different relative humiditiescan usually be omitted. M e r e h ~ d i t y - p r ocontainers t ~ ~ g are not used, appropriate stability data should be provided. c. Accelerated and Stress Conditions

A s previously noted, the expiration dating should be based on ~

- ~ ~ ~ - t edata. r n ~ r a e dthe drug substance and drug Eowever, it is strongly ended that studiesbe ~ n d u ~ on prodw under accelerated andstress conditions. Studies under acceleratedc o n d i ~ n may s provide usekl support data for e s ~ b ~ ithe s ~expiration g date, provide product stability~ o ~ t i oro n fbture product development (e.g., preliminary assessmento f proposed m a n u f a c ~ gchanges such as change in formulation, scale-up), assistin validation o f analytical methods for the stability program, or generate information that may help elucidate d thee ~ t i o profile n of the drug substance or drug product. Studies understress conditions may beuseM in d e t e whether ~ accidental exposuresto conditions otherthan those proposed (e.g., duringt r ~ p o ~ t i oare n) deleterious to the product&nd also for evaluating which specific test parameters maybe the best indicators ofproduct stability. Studieso f the exposure of the drug substance or drug productto extreme conditions may help to reveal patterns of degradation; i f so, such changes shouldbe monitored under proposed storage conditions.Al~oughthe OCH QIA guidance on stability describes the conditionso f the accelerated and stress study7 the applicant should note that those gi~ ~ n ~ t i o should n s be conditions m y not be approp~atefor b i o t e c ~ o l o g i ~ i o l oproducts. carefully selected on a c a s e - b y e e basis. d. Light 78

raft -~ ofort ~ ~ l ~ ~ n t a t i o n 2498 2499 2500

Appli-ts should consult

2501 2502 2503 2504 2505 2506

Changes in the quality o f the product may occur due to the interactions between the formulated Where the lacko f interactions cannotbe ~ o t ~ ~ o l o g i ~ ~product l o l and o g container/closwe. ~ ~ excluded in liquid products (otherthan sealed ampules), stability studies should include samples ~n~~ in the invertedor horizontal position (i.e.,in contact with the closure)as well as in the be upright position,to determine theeffects o f the closure on product quality. Data should supplied for all different ~ n ~ e r / c lcombiitions o s ~ that will be marketed.

2507 2508 2509 25 10 251 l

In addition to the data n e c e s s q for a'conventionalsingle" vial, the ~ p l ishould ~ t demo^ that the closureused with a mdtiple-dose vialis capable o f ~~~g the ~ n ~ to firepeated o ~ insertions and ~ t h so that ~ the~ product s retains its M1 potency, purity, and quality for the maximum period specified in the ~ c t i o n s - f o r - on ~ econtainers, ,andlor package inserts.Such labeliig should be in accordance with FDA requirements.

25 12 2513 25 14 2515 2516

the FDA on a case-by-case basis to determine guidance for testing.

e. ContainerKlosure

f. Stabiiity &er Reconstitution o f Freeze-Dried Product

The stability of fieezedried products after their reconstitution should be demonstrated for the conditions and the m~~ storage period specified on containers, packages, and/or package inserts. Such labeliig should be in accordance with FDA requirements. 7.. Testing Frequency DCH QSC]

25 17 2518 25 19 2520 252l 2522 2523 2524

The shelf lives o f b i o t e c ~ o l o g i ~ ~ i o products l o g i ~ l may vary fiom daysto several years.Thus, it is difficult todraft uniform guidances regardmg the stability study duration and testing fiequency that wouldbe applicable to all types of b i o t ~ ~ o l o g i ~ i o l oproducts. g i ~ l With only a few exceptions, however, theshelf lives for existing products and potentialfitme pmyiucts willbe within the range of 0.5 to 5 years. Therefore, the guidance is based upon expeeteashelf lives in that range. This takes into account the fact that degradation of ~otec~ologi~~iologi~ products may not be governed by the m e factors during different intervals o f a long storage period

2525 2526 2527

2528 2529

When shelf livesofone year or less are proposed, the real-time stabi$ty studies should be conducted monthly for the first three months and at three month intervals therdter. For products with proposed shelf lives of greater than oneyear, the studies shouldbe conducted every three months during the firstyear of storage, everysix months during the secondyear, and ~ u a l l y thereafter.

2530 253 1 2532 2533 2534

While the testing intervals listed above may be appropriate in the preapproval or prelicense stage, reduced testing may be approp~ateafter approval or licensing where dataare available that demonstrate adequate stability.W e r e data exist that indicate the stability of a product is not compromised, the applicantis encouraged to submit a protocol that supports eliination of specific test intervals (e.g., nine-month testing) for~ s ~ p p r o v a ~ p o s t l i ~ long-term e ~ i n g , studies. 79

2535

8. S p e c i ~ ~ t i o [ICH n s Q5CJ

Although b i o t ~ ~ o l o g i ~ i o l o g iproducts cal may be subject to significant losseso f activity, p h y s i c ~ h e ~changes, i ~ l ord e ~ r a ~ t i during on storage, international and national ~ g u l ~ ~ o ~ have provided littleg ~ ~with c respect e to distinct release and end o f shelf life s ~ i ~ ~ t i